A S C : T

A S C : T
A NALYSIS OF S EMANTIC C LASSES : T OWARD N ON -FACTOID
Q UESTION A NSWERING
by
Yun Niu
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy
Graduate Department of Computer Science
University of Toronto
c 2007 by Yun Niu
Copyright Abstract
Analysis of Semantic Classes: Toward Non-Factoid
Question Answering
Yun Niu
Doctor of Philosophy
Graduate Department of Computer Science
University of Toronto
2007
The task of question answering (QA) is to find the accurate and precise answer to a natural
language question in some predefined text. Most existing QA systems handle fact-based questions that usually take named entities as the answers. In this thesis, we focus on a different type
of QA—non-factoid QA (NFQA) to deal with more complex information needs. The goal of
the present study is to propose approaches that tackle important problems in non-factoid QA.
We proposed an approach using semantic class analysis as the organizing principle to answer non-factoid questions. This approach contains four major components:
• Detecting semantic classes in questions and answer sources
• Identifying properties of semantic classes
• Question-answer matching: exploring properties of semantic classes to find relevant
pieces of information
• Constructing answers by merging or synthesizing relevant information using relations
between semantic classes
We investigated NFQA in the context of clinical question answering, and focused on three
semantic classes that correspond to roles in the commonly accepted PICO format of describing
clinical scenarios. The three classes are: the problem of the patient, the intervention used to
treat the problem, and the clinical outcome.
i
We used rule-based approaches to identify clinical outcomes and relations between instances of interventions in sentences.
We identified an important property of semantic classes—their cores. We showed how
cores of interventions, problems, and outcomes in a sentence can be extracted automatically by
developing an approach exploring semi-supervised learning techniques. Another property that
We analyzed is polarity, an inherent property of clinical outcomes. We developed a method
using a supervised learning model to automatically detect polarity of clinical outcomes.
We built explicit connection between text summarization and identifying answer components in NFQA and constructed a summarization system that explores a supervised classification model to extract important sentences for answer construction. We investigated the role of
clinical outcome and their polarity in this task.
ii
Dedication
To my parents.
iii
Acknowledgements
I wish to express my deep appreciation to my advisor, Graeme Hirst, for his help, which
led me into the community of Computational Linguistics, and for his continued encouragement
and support. His valuable guidance aided me immensely during the preparation of this thesis.
I would like to thank my committee members. I thank Gerald Penn and Suzanne Stevenson
for helpful comments and suggestions that led to many improvements in the thesis. I thank
John Mylopoulos and Sharon Straus for being in my thesis committee and their support in the
EPoCare project, and Claire Cardie for being the external reviewer.
I would like to thank Xiaodan Zhu and Jianhua Li for their contributions, Xuming He for
many helpful suggestions, and Afsaneh Fazly for her valuable comments on the writing of the
thesis.
I would also like to thank Patricia Rodriguez Gianolli and Gregory McArthur for many
interesting discussions on the EPoCare project.
I am grateful to have Afra Alishahi, Vivian Tsang, Saif Mohiuddin Mohammad, Faye Baron
and other students in the Computational Linguistics group who made graduate study fun.
I am lucky to have many friends at the University of Toronto whose company and companionship are always there when I need it.
Finally, this thesis would not have been possible without the support and love of my mother,
my father, and Ou. Thank you!
iv
Contents
1
Introduction
1
1.1
What is Question Answering? . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1.2
What are the new research problems posed by QA? . . . . . . . . . . . . . . .
2
1.3
Question-Answering framework . . . . . . . . . . . . . . . . . . . . . . . . .
2
1.4
Fact-based questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
1.4.1
Knowledge-Intensive Approaches . . . . . . . . . . . . . . . . . . . .
5
1.4.2
Data-Intensive Approaches . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.3
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.5
1.6
2
1.5.1
Clinical question answering as NFQA . . . . . . . . . . . . . . . . . . 16
1.5.2
Current research in NFQA . . . . . . . . . . . . . . . . . . . . . . . . 18
Overview of contributions of thesis . . . . . . . . . . . . . . . . . . . . . . . . 22
Our approach for NFQA: semantic class analysis
2.1
3
Non-factoid QA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
26
Our approach of semantic class analysis . . . . . . . . . . . . . . . . . . . . . 26
2.1.1
Representing scenarios using frames . . . . . . . . . . . . . . . . . . . 27
2.1.2
Main components of a QA system guided by semantic class analysis . . 30
2.1.3
The EPoCare Project . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Identifying semantic classes in text:
filling the frame slots
3.1
36
Identifying clinical outcomes using
a combination approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
v
4
5
3.1.1
Detecting clinical outcomes in text . . . . . . . . . . . . . . . . . . . . 37
3.1.2
Determining the textual boundary of clinical outcomes . . . . . . . . . 39
3.2
Analysis of Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Cores of semantic classes
49
4.1
Importance of cores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.2
Architecture of the method . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.3
Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.4
Representing candidates using features . . . . . . . . . . . . . . . . . . . . . . 55
4.5
Data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.6
The model of classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.7
Results and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.7.1
Experiment 1: Evaluation of feature sets . . . . . . . . . . . . . . . . . 63
4.7.2
Experiment 2: Evaluation of candidate sets . . . . . . . . . . . . . . . 65
4.7.3
Experiment 3: Comparison of the semi-supervised model and SVMs . . 67
4.7.4
Experiment 4: Evaluation of distance measures . . . . . . . . . . . . . 70
4.8
Related work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.9
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Polarity of Clinical Outcomes
74
5.1
Related work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2
A supervised approach for clinical outcome detection and polarity classification 77
5.3
5.2.1
Unigrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.2.2
Context features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.2.3
Semantic types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.3.1
Outcome detection and polarity classification in CE text . . . . . . . . 81
5.3.2
Outcome detection and polarity classification in Medline . . . . . . . . 85
5.4
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.5
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
vi
6
Sentence Extraction using Outcome Polarity
6.1
Related work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
6.2
Clinical Evidence as a benchmark . . . . . . . . . . . . . . . . . . . . . . . . 94
6.3
Identifying important sentences . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.3.1
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.3.2
Features to identify important sentences . . . . . . . . . . . . . . . . . 96
6.4
Data Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
6.5
Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.6
7
92
6.5.1
Sentence-level evaluation . . . . . . . . . . . . . . . . . . . . . . . . . 98
6.5.2
ROUGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Conclusion
105
7.1
Summary of contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.2
Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.2.1
Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
7.2.2
Directions for future work . . . . . . . . . . . . . . . . . . . . . . . . 108
Appendices
113
A List of abbreviations
113
B A subset of syntactic tags in Apple Pie Parser
114
C The effect of σ in the RBF kernel in core identification
115
D Algorithm for Boundary Detection of Clinical Outcomes
116
E Sample output of MetaMap
118
F Sample output of Minipar
122
G List of words for building CHANGE PHRASES features
124
vii
H Using an RBF kernel in detecting polarity of clinical outcomes
I
J
127
Results of the summarization with different feature sets in sentence-level evaluation
128
F-score of combinations of features in each single summary
129
Bibliography
131
viii
List of Tables
2.1
The treatment frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.1
Results of identifying outcomes in CE . . . . . . . . . . . . . . . . . . . . . . 38
3.2
Results of boundary detection of correctly identified outcomes in CE . . . . . . 42
3.3
Cue words/symbols for relations between interventions . . . . . . . . . . . . . 47
3.4
Results of relation analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1
Number of Instances of Cores in the Whole Data Set . . . . . . . . . . . . . . 59
4.2
Number of Instances of Target Classes in the Candidate Set . . . . . . . . . . . 64
4.3
Number of Candidates in Different Candidate Sets . . . . . . . . . . . . . . . . 66
4.4
Results of Classification on Different Candidate Sets . . . . . . . . . . . . . . 67
4.5
F-score of Classification Using Different Models . . . . . . . . . . . . . . . . 70
4.6
Accuracy using different distance measures. . . . . . . . . . . . . . . . . . . . 70
5.1
Accuracy of positive/negative classification using a linear kernel in CE . . . . . 82
5.2
Number of instances in each class (CE) . . . . . . . . . . . . . . . . . . . . . 83
5.3
Results of the four-way classification with different feature sets in CE . . . . . 84
5.4
Classification results of each class on CE data . . . . . . . . . . . . . . . . . . 85
5.5
Number of instances in each class (Medline) . . . . . . . . . . . . . . . . . . . 86
5.6
Results of two-way and four-way classification with different feature sets (Medline) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.1
F-score of the summarization with different feature sets in sentence-level evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6.2
Comparison of feature sets in single summary at compression ratio 0.25 . . . . 102
ix
6.3
ROUGE-L score of different feature sets . . . . . . . . . . . . . . . . . . . . . 103
H.1 Results of positive/negative classification using an RBF kernel . . . . . . . . . 127
I.1
Results of the summarization with different feature sets in sentence-level evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
x
List of Figures
1.1
QA Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
1.2
A subset of the answer-type taxonomy [Paşca and Harabagiu, 2001b] . . . . . .
6
1.3
Semantic forms of sentences ([Paşca and Harabagiu, 2001b]) . . . . . . . . . .
9
1.4
Example of a clinical question, with corresponding evidence from Clinical Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1
EPoCare system architecture. . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2
Disease categories in Clinical Evidence . . . . . . . . . . . . . . . . . . . . . 33
2.3
Our work in the QA framework . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1
Cue words for detecting clinical outcomes . . . . . . . . . . . . . . . . . . . . 37
3.2
Examples of output of Apple Pie Parser . . . . . . . . . . . . . . . . . . . . . 40
3.3
An example of cue word difference . . . . . . . . . . . . . . . . . . . . . . . . 41
4.1
Architecture of the approach of core identification . . . . . . . . . . . . . . . . 52
4.2
Example of output of MetaMap . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3
Example of dependency triples extracted from output of Minipar parser. . . . . 58
4.4
Manifold structure of data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.5
Classification Results of Candidates . . . . . . . . . . . . . . . . . . . . . . . 64
4.6
Linear separating hyperplanes in two dimensions. The support vectors are
marked by squares. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
5.1
Accuracy of classification using different fractions of training data . . . . . . . 83
6.1
Comparison of features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
xi
C.1 Classification results with different values for σ . . . . . . . . . . . . . . . . . 115
F.1
Example of dependency triples extracted from output of Minipar parser. . . . . 123
J.1
The performance of different combinations of features in each summary . . . . 130
xii
Chapter 1
Introduction
1.1
What is Question Answering?
As more and more information is accessible to users, more support from advanced technologies
is required to help obtain the desired information. This brings new challenges to the area of
information retrieval (IR) in both the query and the answer processing. To free the user from
constructing a complicated boolean keywords query, the system should be able to process
queries represented in natural language. Instead of replying with some documents relevant to
the query, the system should answer the questions accurately and concisely. Systems with such
characteristics are Question-Answering (QA) systems, which take advantage of high-quality
natural language processing and mature technologies in IR. The task of a QA system is to find
the answer to a particular natural language question in some predefined text.
Generally, current QA tasks can be classified into two categories: fact-based QA (FBQA)
and non-factoid QA (NFQA). In FBQA, answers are usually named entities, such as person
name, time, and location. For example:
Q: Who was the US president in 1999?
A: Bill Clinton
Q: Which city is the capital of China?
A: Beijing
NFQA aims to answer questions whose answers are not just named entities, such as questions
1
C HAPTER 1. I NTRODUCTION
2
posed by clinicians in patient treatment:
Q: In a patient with a generalized anxiety disorder, does cognitive behaviour or relaxation therapy
decrease symptoms?
Clinical outcomes of cognitive behaviour or relaxation therapy could be complicated. They
could be beneficial or harmful; they could have different effects for different patient groups;
some clinical trials may show they are beneficial while others don’t. Answers to these questions
can only be obtained by synthesizing relevant information.
Both FBQA and NFQA need to address some major research problems, and they fit into
the same general QA framework. This thesis focuses on NFQA, and our working domain is
medicine.
1.2
What are the new research problems posed by QA?
The first problem for QA is to understand the task. Since there are many different types of
questions, it is very important for a QA system to know what a particular question is asking
for. Some techniques have been recognized to be effective in FBQA, which are discussed in
section 1.4.1 on the question-processing phase. In NFQA, however, it is much more difficult
to understand the information needs.
Matching the answer to the question is another big challenge. Questions and the answers
often have very different phrasings. Matching techniques need to find the correspondence
between them. Compared to FBQA, such correspondence in NFQA is usually less explicit.
Answer generation is the last problem in QA. After the best candidates are selected by the
matching techniques, they need to be processed to obtain accurate and concise answers.
1.3
Question-Answering framework
The architecture of a typical QA system is shown in Figure 1.1.
1. Question processing. The aim of the question processing is to understand the question.
In most FBQA systems, this includes:
C HAPTER 1. I NTRODUCTION
3
Question
Set
Question
Processing
Q-A Matching
Answer Generation
Document
Processing
Document
Collection
Figure 1.1: QA Architecture
• determining what type of question is being asked (e.g., where)
• inferring what kind of answer is expected (e.g., location)
• determining the focus of the question—its central point
• formulating a query on the document collection by using keywords in the question
Some NFQA systems suggest to clarify the question by communicating with users.
2. Document processing. Before the matching of question and answer starts, the documents in the collection may be transformed to some other representation so that efficient search can be performed. Many systems imported index technology from IR
systems in this step.
3. Question-Answer matching. Before doing detailed analysis to find the answer, a relatively small set of candidates should be found. Conventional keyword matching and
expected answer-type checking are often involved in this step. Unmatched candidates
C HAPTER 1. I NTRODUCTION
4
will be filtered out directly.
To find the best answer, different techniques are used to analyze the relationship between the question and candidate answer thoroughly. Knowledge-intensive, dataintensive, and statistical approaches address the problem with different emphasis.
4. Answer generation. It has not been fully addressed in most current systems. Most systems just extract small fragments that contain the answer information from the answer
candidates as the final answers. Even the extraction process is not discussed in detail
in many works.
In the following two sections, related work in FBQA and NFQA is reviewed respectively
to further understand their difference and connections, and state-of-the-art in QA.
1.4
Fact-based questions
The main problem in QA is the great variation in expressing the question and the answer.
According to how it is addressed, current work in FBQA can be partitioned into two classes.
• Knowledge intensive. The intuition of the knowledge-intensive approaches is to find
a proper meta-form so that both the question and the answer can be represented by it.
The construction of the form usually exploits natural language processing technology
as well as related real-world knowledge.
• Data intensive. The data-intensive approaches put the emphasis on prediction of the
answer by using the evidence from the data set. For each question, some approaches
try to compose all the possible answer formats and then compare them with the answer
candidates to find the one that meets the prediction. Some approaches estimate how
likely a candidate is the expected answer by collecting statistical data from a large
candidate set.
The following two subsections will discuss some typical FBQA work in detail.
C HAPTER 1. I NTRODUCTION
5
1.4.1 Knowledge-Intensive Approaches
The problems that are emphasized in knowledge-intensive systems are discussed in this section.
For each problem, methods explored in different systems are compared.
Answer-type identification The type of the answer tells us the general category of the expected answer, whether it is a person, a location, or a time etc. To determine the answer type,
the type of the question should be identified first. As mentioned earlier, knowing the question
type addresses the “what to find” problem. Since most FBQA systems focus on wh- questions
(who, when, where, why, what), it is natural to classify the types according to the stem of the
question: the wh- words.
Most answers for wh- questions are related to named entities (NE); thus most FBQA systems classify the answers by different types of NE, such as: time, product, organization, person,
etc. The NE identification technique from information extraction (IE) is quite helpful and usually is imported into this process. There are some other answer categories that do not belong
to NE. As in Paşca’s work [Paşca and Harabagiu, 2001b], type reason is applied to the why
questions and type definition is included for questions asking for the definition of a concept.
A parser is often involved to find the answer type. For example, in Paşca’s work, it depends
on a concept hierarchy “Answer-Type Taxonomy” and a special-purpose parser.
• Answer-type taxonomy. The taxonomy is a tree structure constructed off-line which
contains all the answer types that can be processed by the system. It is built in a topdown manner with general concepts on the top and more-specific concepts on lower
levels. A subset of the taxonomy is shown in Figure 1.2.
The top level of the hierarchy contains the most representative conceptual nodes, e.g.
person, location, money, nationality, etc. Some of these are further categorized
to more specific concepts. For instance, the location node is divided into university, city, country, etc. Some concepts are connected to corresponding synsets from
WordNet. As an example, person is linked to several sub-trees rooted separately at
scientist, performer, European, etc. It is worth noticing that although most concepts in the taxonomy are nouns, there are verbs and adjectives as well. A concept can
C HAPTER 1. I NTRODUCTION
6
Top
time
product
money
university
location
country
city
mammal
nationality
person
numerical value landmark
other_loc
province
degree
duration
dimension
Figure 1.2: A subset of the answer-type taxonomy [Paşca and Harabagiu, 2001b]
be connected to the WordNet noun sub-hierarchies, verb sub-hierarchies or, adjectival
satellites. For example, the product node is connected to nouns {artifact, artefact}
and verbs {manufacture, fabricate, construct}. The whole taxonomy was constructed
manually and was adapted to the sample questions.
• Parser. The wh- word in a question cannot always provide enough information on
the type of the expected answer. For example, what can ask for many different types
of things. To solve this problem, Paşca implemented a parser to find the word(s) in
a question that help determine the expected answer type. The question is parsed to
locate the word(s) that has a head-modifier dependency to the question stem (the whword). For instance, in the question What do people usually buy in Hong Kong?, what
is a dependent of the verb buy. The word buy is then mapped onto the answer-type
taxonomy to obtain the expected answer type, e.g. product.
This method is effective in determining the answer type at above 90% accuracy in the TREC
test questions. However, there is a lot of manual work involved. Currently, the answer-type
taxonomy encodes 8707 English concepts with 153 connections to WordNet sub-hierarchies
[Paşca and Harabagiu, 2001b]. It will be quite burdensome to include more and more nodes
into the taxonomy for the system to be adaptive to new expressions of questions and answers.
The approach explored by [Hovy et al., 2000] is similar to Paşca’s work. In their “Webclopedia” system, they also built a taxonomy of answer types (“QA Typology”) but WordNet
C HAPTER 1. I NTRODUCTION
7
is not involved. The typology contains 94 nodes [Hovy et al., 2000]. An extended parser is
also used in the process of answer-type identification, which contains some semantic background knowledge. A set of manually constructed rules is included in the parser to determine
the correct answer type. The answer type produced by the parser can be the concepts in the
QA Typology, P-O-S tags, roles produced in the parse tree, or concepts from the semantic type
ontology of the parser.
Identification of question (answer) focus As defined by [Moldovan et al., 1999], “a focus
is a word or a sequence of words which define the question and disambiguate it in the sense
that it indicates what the question is looking for, or what the question is all about” (page
176). For example, the question What type of bridge is the Golden Gate Bridge? [Paşca and
Harabagiu, 2001a] has bridge as the answer type and type as the answer focus. From the
definition, the focus is very important for answering a question. Some systems mentioned
the concept explicitly [Moldovan et al., 1999; Harabagiu et al., 2000; Ferret et al., 2001; Lee
et al., 2001], others may include it in the general answer type identification process without
discussing it separately. In both cases, no method or technique is provided to address the
problem particularly.
Query generation The query-generation process usually involves keyword extraction from
the original question with or without weights attached to them. For example, a query corresponding to the question Who invented the paper clip? is [paper AND clip AND invented].
Later, the query can be expanded by using a knowledge base such as WordNet. In some systems, after removing stop words, the keywords are selected by a set of heuristics [Moldovan
and Harabagiu, 2000; Lee et al., 2001; Alpha et al., 2001]. Different systems have different
preferences in whether lemmata or stemmed words should be used as keywords.
Query expansion is often applied to make sure that the correct answer will not be missed.
Most systems use synonyms of the selected keywords in WordNet to expand the query. More
sophisticated query-expansion techniques are explored in Harabagiu’s work [Harabagiu et al.,
2001a], which include three levels of alternations:
• Morphological alternations. When no answer is found by matching the original key-
C HAPTER 1. I NTRODUCTION
8
words from the question, the morphological alternations are considered. For example,
the noun inventor will be added to the query because of the original verb invent.
• Lexical alternations. WordNet is a source for adding lexical alternations to the query.
In most cases, synonyms of a word are added, although other relationships may also
be considered. For example, killer has a synonym assassin which should be included
in the query expansion.
• Semantic alternations. The semantic alternations are defined as “the words or collocations from WordNet that (a) are not members of any WordNet synsets containing the
original keyword; and (b) have a chain of WordNet relations or bigram relations that
connect it to the original keyword” [Harabagiu et al., 2001a] (page 278). For example,
the candidate words can be hypernyms or hyponyms of the original word, or even just
related to it in some situation. To answer the question How many dogs pull a sled in
the Iditarod?, since sled and cart are found to be forms of vehicles, the word harness
that is related to pull cart is included in the query expansion.
Three heuristics are constructed to decide when and how to perform these alternations.
However, for the semantic alternations, the heuristic does not specify which semantic relations
should be considered in a particular situation (in fact, it is almost impossible to do so). This
kind of problem seems to be an inherent limitation of knowledge-based approaches.
Matching of Question and Answer Keyword matching is the first criterion to filter out
irrelevant answers in almost all FBQA systems. To make the system efficient, usually only text
fragments that contain the query keywords will be returned instead of the whole documents.
The number of returned fragments is fairly large, although it varies in different systems. The
query keywords are expanded or shrunk to make sure that the proper number of fragments are
returned.
In the next step, further matching is performed. Fragments that do not meet the strict
requirements are filtered out. The filter can be used to verify the semantic relations or can
be used as some ranking scheme. In Harabagiu’s work, the filter is executed at three levels
C HAPTER 1. I NTRODUCTION
9
[Harabagiu et al., 2001a]. At the first level, fragments that do not contain at least one concept
of the same semantic category as the expected answer type are filtered out.
At the second level, the question represented in its semantic form produced by the parser in
the question-processing phase is unified with the semantic forms of the fragments that contain
the possible answers (answer candidates). The aim of the unification is to check how much information contained in the query is also contained in the answer candidates. Thus, the question
concepts as well as the dependencies of the query terms which are represented by the semantic
form are compared with the semantic forms of the answer candidates.
The semantic form of a sentence is derived from its syntactic parse tree. To construct the
semantic form, the semantic concept that the sentence is about (the answer type) is added to
the tree (which works as a slot in the question representation and the slot filler in the answer
representation). Unimportant words are removed. Figure 1.3 is an example of the semantic
forms:
Question:
What company sells most greeting cards?
ORGANIZATION sells greeting cards most
Answer:
Hallmark remains the largest maker of greeting cards
maker
greeting cards
largest
ORGANIZATION(Hallmark)
Figure 1.3: Semantic forms of sentences ([Paşca and Harabagiu, 2001b])
At the third level, the question and answer candidates are represented in their logical forms.
The logical relations held by the terms in the query are evaluated in the abduction of answers.
As an example of the logical form, the question Why did Hong Li bring an umbrella? can be
represented as:
C HAPTER 1. I NTRODUCTION
10
[REASON (x)&Hong(y)&Li(y)&bring(e, x, y, z)&umbrella(z)]
In this example, Hong and Li are identified as the same entity by using the same symbol y.
A candidate is selected to be an answer if it can be proved by using the logical form. The
prover processes the terms in the query logical form from the left to the right. For each term,
it tries to identify corresponding information contained in the answer logical form. Real-world
knowledge is needed here. For example, in the above question it may be helpful to know that
Hong Li is a person’s name.
In the work of [Hovy et al., 2001], answer type and answer focus are checked in the parse
trees of the question and candidate answers in the matching process. When it is not enough
for selecting a good answer, several heuristics are applied which consider the expected answer
range, knowledge of abbreviation, and knowledge of some formats of special information (e.g.
e-mail address, post code) etc.
Some systems try to choose the answer by ranking the candidate answers. The ranking
often depends on heuristic rules about how the answer candidates contain the query terms,
e.g., the order of the query terms in the answer candidates, the number of query terms that are
matched, the distance from the position of the embedded answer type to the query terms etc.
Some systems [Ferret et al., 2001; Prager et al., 2000; Srihari and Li, 1999; Alpha et al., 2001]
implement the matching by ranking the passages according to weighted features or terms which
are chosen off-line. Machine learning techniques are imported in the ranking in some systems.
Paşca and Harabagiu [2001b] use a perceptron model to compare two candidates, while Prager
et al. [2000] apply logistic regression to score NEs contained in the candidates.
Answer Extraction The task is to extract a concise answer from the answer candidates.
In some systems, the candidates are strings in text windows of specified size [Moldovan et al.,
1999]. Some others consider sentences as the candidates [Ferret et al., 2001; Hovy et al., 2001].
The candidates with the highest scores obtained in the matching process are extracted as the
final answers. In knowledge-based systems, no particular techniques are applied to extract the
answer.
C HAPTER 1. I NTRODUCTION
Evaluation
11
In the TREC-10 evaluation, the LCC system developed by Harabagiu et al.
[2001b]performs well in both the main task and the list question task with the mean reciprocal rank (MRR)1 of 0.57 in the main task and accuracy (no. of distinct instances/target no. of
instances) of 0.76 in the list task. LCC is ranked second in the main task and first in the list
task. The work of [Hovy et al., 2001] and [Alpha et al., 2001] are also in the top five ranked
systems.
1.4.2 Data-Intensive Approaches
IE-based QA
The NE identification techniques from IE are exploited in most FBQA sys-
tems. Some systems demonstrated that other techniques may also be helpful in QA. In the
TREC-10 main task, the system that had the best performance was the one from InsightSoft-M
[Soubbotin, 2001]. It applies a set of pre-defined patterns generated by analyzing the document
collection to the answer candidates to match particular types of questions.
The idea is similar to pattern-matching and slot-filling in IE. Since various tasks should be
addressed in a QA system, to take advantage of the pattern-matching technique, the Insight
system classifies the questions into different categories and then constructs patterns for each
category. There are two categories of patterns in the system:
• patterns representing a complete structure
Example:
capitalized word; parenthesis; four digits; dash; four digits; parenthesis
would match “Mozart (1756 - 1791)”
• patterns composed by specific pattern elements
Example:
[number]+[term from currency list]
would match “5 cents”
In the system, questions must be analyzed to obtain the accurate answer type so that the
correct patterns will be triggered later. Relevant passages are obtained by searching for query
1
MRR is an accuracy measure. To calculate the MRR, “an individual question receives a score equal to the
reciprocal of the rank at which the first correct response is returned, or zero if none of the five responses contained
a correct answer” [Voorhees, 2001].
C HAPTER 1. I NTRODUCTION
12
keywords in the document collection. Only these passages are compared with the patterns to
identify potential answers. As the patterns should adapt to various phrasing of answers, a great
number of patterns are constructed manually for each question type (e.g., 23 patterns are built
for the who-author type of questions). As in IE systems, manually constructing patterns is a
very time-consuming task. So automatic pattern construction may be the future work of the
pattern-matching-based systems.
As shown in the above examples, the system contains a set of pattern elements such as
currency, person names, country names, etc. It seems that if an NE recognizer is used to
replace these elements, the patterns can be simpler.
We can see that almost no deep knowledge analysis is involved in the pattern construction.
However, proper knowledge is necessary in the system to process complicated questions. For
example, the ambiguous question Who is Bill Gates? is actually asking for the reason why
Bill Gates is famous. As in the knowledge-intensive approaches, the identification of the correct answer type is important in finding the correct answer here. It is declared by [Soubbotin,
2001] that detailed categorization of question types is a precondition for effective use of the
method.
Since it is not possible to construct complete patterns for a type of question, for those cases
in which the questions do not match any patterns, the system tries to select the answers by
comparing lexical similarity of the question and answer candidates.
Another IE-based system was described in [Srihari and Li, 1999]. The idea is to answer
questions by executing IE in three levels:
• Named entity: extract named entities as answer candidates
• Correlated entity: extract pre-defined relationships between the entities
• General events: extract different roles in general events (e.g. who did what to whom
when and where)
However, only the first level was completed in the system.
From the above analysis we can see that IE-based systems lie somewhere between the
knowledge-intensive and the data-intensive methods. The advantage of them is that the burden
C HAPTER 1. I NTRODUCTION
13
of deriving and comparing the semantic and logic similarity of question and answer is released
to some degree.
Redundancy-based QA
The two other systems that also imported pattern-matching tech-
niques into the query answer-matching process are MultiText [Clarke et al., 2000] and AskMSR
[Dumais et al., 2002]. In MultiText, the patterns used consist of regular expressions with simple
hand-coded extensions.
AskMSR just applies simple string-based manipulations in the query rewriting to formulate
the patterns. The rewrite rule is a triple of the form [string, L/R/-, weight], where “string” is
the reformulated search query, “L/R/-” indicates the position in the text where the answer is
expected to find with respect to the query string (left, right or anywhere) and “weight” is a
confidence figure for a particular query. If a query pattern is more likely to find the correct
answer, it will have a higher weight than others. The following is an example [Dumais et al.,
2002]:
Question: Who created the character of Scrooge?
Rewrite1: [created + the character + of Scrooge, left, 5]
Rewrite2: [+the character + of Scrooge + was created + by, right, 5]
However, the matching process is not the only component that helps find the answer in the
two systems. The redundancy of the data is further explored to obtain the answer. The idea that
data redundancy can be applied to question answering is basically the same in the two systems
with slight differences. As indicated by [Clarke et al., 2001], the hypothesis was that correct
answers could be distinguished from other candidates solely by their repeated occurrence in
relevant passages.
The hypothesis is implemented by assigning weights to the candidate answers. In MultiText, after the pattern matching, the retrieved answer candidates are ranked according to the
sum of the weights of the candidate answer terms that they contain. To calculate the term
weights, [Clarke et al., 2001] used an idf -like formula with a redundancy parameter. The redundancy parameter is defined as the number of retrieved passages in which a particular term
appears. In the answer generation process, the segment in a passage that maximizes the sum of
the term weights it contains is extracted. MultiText ranked among the top five systems in the
TREC-10 main task and list task evaluation.
C HAPTER 1. I NTRODUCTION
14
In AskMSR, the n-grams (1-, 2-, 3- grams) in the retrieved passages are the candidates
to be ranked. The weight of an n-gram depends on the confidence value of the rewrite rules
that generated it (“5” in the above example of rewrite rules). The confidence values in all the
unique retrieved passages in which the n-gram occurred are summed up to obtain the score of
the n-gram. The n-grams are then filtered and re-weighted by a set of manually constructed
heuristics. Finally, the remaining ones are tiled to get the answer. Tiling forms longer n-grams
by merging overlapping shorter n-grams. For example, “A B C” and “B C D” is tiled into “A B
C D.” Compared with MultiText, the system does not need the corpus to be full-text indexed,
nor does it need global term weights. However, AskMSR performs worse then MultiText. It
was not in the top eight systems in the TREC-10 evaluation.
Although AskMSR was not so successful as MultiText in TREC-10, the idea that redundancy can help find the answer by using only simple patterns is verified by Dumais et al.
[2002]. Their results show that the system performs much better on Web data than on the
TREC data (the former is much larger than the latter). It is the same in MultiText. Compared
with other approaches, a major contribution of redundancy-based methods is that they explore
the relationships among good answer candidates. The correct answer of a question may be
very difficult to identify because of its complicated formulation. However, it may be promoted
by many relevant answer candidates that have simpler phrasings. This important information
is ignored in other approaches.
Statistical QA
Not many systems explored statistical approaches for FBQA. This might be
because the potential of statistical models had not been realized in late 90’s. Among the topranked systems in TREC-10, only one system [Ittycheriah et al., 2001] included a statistical
model.
The system architecture in [Ittycheriah et al., 2001] is similar to the general architecture described in section 1.3. The statistical model is not applied to the whole system but rather on two
components of the system: answer-type prediction and answer selection. The NE recognition
is also implemented by using the statistical method. All three tasks are viewed as classification
problems and the maximum-entropy models are constructed with three different feature sets.
The answer types include the standard categories of NE in the Message Understanding Con-
C HAPTER 1. I NTRODUCTION
15
ference (MUC) plus two more types: reason for why questions and phrase for all the others.
The features take care of unigrams, bigrams, PoS, the position of the question words, as well
as some expansion in WordNet. In answer selection, 31 features related to sentence, entity,
definition, and linguistics (e.g. the answer candidate is either in the subject or object position
etc.) are constructed. The NE annotation cares about the words, morphs, PoS, and grammar
flags.
For the answer-type classification, 3300 questions were annotated manually before training.
The training set for the answer-selection task is 400 question–answer pairs from TREC-8 and
TREC-9. Because of the availability of the training data for the categories of NE, the answer
types are almost confined to the MUC classes. The difficulty of obtaining enough training data
is one problem that affects the system performance.
The statistical model works well in the answer-type identification task (accuracy 90.5%).
The results for NE recognition are not reported in the paper, although it is indicated by error
analysis that the performance is good. This system is one of the top-ranked systems in TREC10, which indicates the effectiveness of using statistical models in FBQA.
1.4.3 Summary
Data-intensive approaches try to answer questions without deeply understanding the meaning
of the questions and the answer text. This reduces the complexity of the system model. However, as we see from the above discussion, a pure data-based method is not enough to construct
a system with high accuracy, because proper knowledge plays an important role in question
analysis, which forms the guide in answer searching.
1.5
Non-factoid QA
In comparison to FBQA, NFQA is much less understood by researchers. However, it is such an
important area that it is attracting more and more research interest [Niu et al., 2003; Diekema
et al., 2003; Stoyanov et al., 2005; DUC, 2005].
NFQA deals with more complex information needs. We observe two distinct characteristics
of NFQA as compared to FBQA.
C HAPTER 1. I NTRODUCTION
16
• Non-factoid questions usually cannot be answered using a word or phrase, such as
named entities. Instead, answers to these questions are much more complex, and often
consist of multiple pieces of information from multiple sources.
• Compared to FBQA, in which an answer can be judged as true or false, NFQA needs
to determine what information is relevant in answer construction.
Some examples of non-factoid questions are as follows.
In a patient with a generalized anxiety disorder, does cognitive behaviour or relaxation therapy
decrease symptoms?
Was the most recent presidential election in Zimbabwe regarded as a fair election? [Stoyanov et al.,
2005]
What advantages/disadvantages does an Aluminum alloy have over Ti alloy as the core for a honeycomb design? [Diekema et al., 2003]
Symptoms in the first question is a general concept, any clinical outcome of cognitive behaviour
or relaxation therapy in anxiety disorder could be relevant. These outcomes could be different
for different patient groups (e.g. different age groups); they may be positive in some clinical
trials while negative in some others. All this evidence should be taken into account in constructing the answer. For the second question, it is not easy to reach an answer of yes or no.
In fact, it might not be possible to do so, as it is very likely that both answers have supporters.
Neither of them should be ignored in the answer. In addition, to either positive or negative attitude, information describing the reasons can be highly desirable. To answer the third question,
we need to synthesize information on various aspects that the two metals are compared.
Because of the complex answers, current FBQA techniques will have difficulty in answering non-factual questions. Therefore, it is important to develop new strategies and techniques
to address new challenges in NFQA.
1.5.1 Clinical question answering as NFQA
Clinicians often need to consult literature on the latest information in patient care, such as side
effects of a medication, symptoms of a disease, or time constraints in the use of a medication.
C HAPTER 1. I NTRODUCTION
17
The published medical literature is an important source to help clinicians make decisions in
patient treatment [Sackett and Straus, 1998; Straus and Sackett, 1999]. Studies have shown
that searching the literature can help clinicians answer questions regarding patient treatment
[Gorman et al., 1994; Cimino, 1996; Mendonça et al., 2001]. It has also been found that if
high-quality evidence is available in this way at the point of care—e.g., the patient’s bedside—
clinicians will use it in their decision making, and it frequently results in additional or changed
decisions [Sackett and Straus, 1998; Straus and Sackett, 1999]. The practice of using the
current best evidence to help clinicians in making decisions on the treatment of individual
patients is called evidence-based medicine (EBM).
Questions posed by clinicians in patient treatment present interesting challenges to an
NFQA system. For a clinical question, it is often the case that more than one clinical trial
with different experimental settings will have been performed. Results of each trial provide
some evidence on the problem. To answer such a question, all this evidence needs to be taken
into account, as there may be duplicate evidence, partially agreed-on evidence, or even contradictions. A complete answer can be obtained only by synthesizing these multiple pieces of
evidence, as shown in Figure 1.4. In our work, we take EBM as an example to investigate
NFQA. Our targets are questions posed by physicians in patient treatment.
Clinical question: Are calcium channel blockers effective in reducing mortality in acute
myocardial infarction patients?
Evidence1: . . . calcium channel blockers do not reduce mortality, . . . may increase mortality.
Evidence2: . . . verapamil versus placebo . . . had no significant effect on mortality.
Evidence3: . . . diltiazem significantly increased death or reinfarction.
Evidence4: . . . investigating the use of calcium channel blockers found a non-significant
increase in mortality of about 4% and 6%.
Figure 1.4: Example of a clinical question, with corresponding evidence from Clinical Evidence.
C HAPTER 1. I NTRODUCTION
18
1.5.2 Current research in NFQA
Unlike FBQA, in which the main research focuses on wh- questions (e.g. when, where, who) in
a rather general domain, most work in NFQA starts with a specific domain, such as terrorism, or
a specific type of question, such as opinion-related questions. The complexity of NFQA tasks
may account for this difference. In this section, current work in NFQA is reviewed according
to different research problems of the QA task that it addresses.
Question processing Because the information needs are more complex, some work put more
efforts on understanding questions. Hickl et al. [2004], Small et al. [2004] and Diekema et al.
[2003] suggest answering questions in an interactive way to clarify questions step by step. In
addition, Hickl et al. argue that decomposition of complex scenarios into simple questions is
necessary in an interactive system. As an example, the complex question What is the current
status of India’s Prithvi ballistic missile project? is decomposed into the following questions
[Hickl et al., 2004]:
1. How should ‘India’ be identified?
2. Pre-independence or post-independence, post-colonial, or post-1947 India?
3. What is ‘Prithvi’?
4. What does Prithvi mean?
5. What class of missiles does Prithvi belong to?
6. What is its range/payload, and other technical details?
7. ...
They propose two approaches to the decomposition: by approximating the domain-specific
knowledge for a particular set of domains, and by identifying the decomposition strategies
employed by human users. Preliminary results from two dialog pilot experiments suggest five
strategies for question decomposition employed by experts that could be helpful in automatic
decomposing complex questions.
C HAPTER 1. I NTRODUCTION
19
Following that work, Harabagiu et al. [2004] derived intentional structure and the implicatures enabled by it for decomposing of complex questions, such as What kind of assistance has
North Korea received from the USSR/Russia for its missile program? The authors claim that
intentions that the user associate with the question may express a set of intended questions; and
each intended question may be expressed as implied questions. The intended questions of this
example include What is the USSR/Russia? What is assistance? What are the missiles in the
North Korean inventory? Then, these intended questions further have implied questions, such
as Is this the Soviet/Russian government? Does it include private firms, state-owned firms,
educational institutions, and individuals? Is it the training of personnel? What was the development timeline of the missiles? Questions like Will Prime Minister Mori survive the crisis?
and Does Iraq have biological weapons? are also questions that this paper is interested in
[Harabagiu et al., 2004].
Two methods of generating the intentional structure of questions are explained by two examples in the paper. One is based on lexico-semantic knowledge bases (e.g. WordNet), and
the other uses the predicate-argument structures of questions. The authors claim that the intentional structure may determine a different interpretation of the question, and answer extraction
depends on the semantic relations between the coerced interpretations of predicates and arguments, although no details of evaluation are described in the paper.
The system HITIQA (High-Quality Interactive Question Answering) [Small et al., 2004]
also emphasizes interaction with user to understand the information needs, although it does not
attempt to decompose questions. During the interaction, the system asks questions to confirm
the user’s needs. After receiving yes or no from the user, the goal of searching is clearer. The
interaction is data-driven in that questions asked by the system are motivated from previous
results of information searching (which form the answer space).
Diekema et al. also suggest to have a question negotiation process for complex QA [Diekema
et al., 2003]. The QA system deals with real-time questions related to “Reusable Launch Vehicles”. For example, broad-coverage questions like How does the shuttle fly?, and questions
about comparison of two elements such as What advantages/disadvantages does an Aluminum
alloy have over Ti alloy as the core for a honeycomb design? are typical in the domain. A
question-answering system architecture with a module of question negotiation between the
C HAPTER 1. I NTRODUCTION
20
system and the questioner is proposed in the paper.
Matching of question and answer Berger et al. [2000] describe several interesting models
to find the connection between question terms and answer terms.
• tf · idf . This model is different from the standard tf · idf calculation. The conventional
IR vector space model is applied in QA by taking the question and answer as different
documents.
Given an m-word question q = {q1 , q2 , ..., qm }, and an n-word answer a = {a1 , a2 , ..., an },
the adapted cosine similarity between the question and the answer is given by the following formula [Berger et al., 2000]:
λ2w · fq (w) · fa (w)
,
score(q, a) = qP
P
2·
2
f
(w)
f
(w)
q
a
w∈q
w∈a
P
w∈q,a
(1.1)
where
λw = idf (w) = log
|D|
|{d ∈ D : fd (w) > 0}|
.
(1.2)
where fd (w) is the number of times word w appears in document d. Here a document
is an answer; D is the entire set of answers.
• Mutual information for query expansion. Instead of searching for terms for query
expansion in a large knowledge taxonomy such as WordNet, a model for calculating
the mutual information of query terms and answer terms is built. In this model, the
mutual information of any pair of terms appearing in the training set of paired questions
and answers is calculated. This can be used to locate the most relevant terms in the
answer that are correlated with any question term. These terms are expected to be good
candidates for expanding the query.
• Statistical translation model. Taking a machine translation view of the QA problem,
the question and answer can be treated as two different languages. The model is built
to learn how an answer a corresponds to a question q by calculating p(q|a).
C HAPTER 1. I NTRODUCTION
21
As indicated by [Berger et al., 2000], these models are presented for a problem slightly
different from a typical QA task. It is to find answers within a large collection of candidate
responses. The responses are supposed to be correct answers to the questions. It is not mentioned in the paper if there are only one-to-one relations between questions and answers. Since
the answers are created according to the questions, it may be the case that the phrasing of
question and answer has more overlap than it does in the general QA task. Also, answers to
different questions may be easier to distinguish. Although such differences exist, the essence
of the answer-finding task and the QA task are the same. Models explored by the former may
adapt to the latter as well. Soricut and Brill [2006] extend Berger’s work to answer FAQ-like
questions. In their work, although FAQ question and answer pairs are used as training data, the
goal is to extract answers from documents on the web, instead of pairing up existing questions
and answers in FAQ corpora. Taking questions and answers as two different languages, a machine translation model is applied in the answer extraction module to extract three sentences
that maximize the probability p(q | a) (q is the question and a is the answer) from the retrieved
documents as the answer.
In system HITIQA, frame structure is used to represent the text, where each frame has
some attributes. For example, a general frame has frame type, topic, and organization. During
the processing, frames will be instantiated by corresponding named entities in the text. In
answer generation, text in the answer space is scored by comparing their frame structures with
the corresponding goal structures generated by the system according to the question. Answers
consist of text passages from which the zero conflict frames are derived. The correctness of the
answers were not evaluated directly. Instead, the system was evaluated by how effective it is in
helping users to achieve their information goal. The results of a three-day evaluation workshop
validated the overall approach.
Cardie et al. [2003] aims to answer questions about opinions (multi-perspective QA), such
as: Was the most recent presidential election in Zimbabwe regarded as a fair election?, What
was the world-wide reaction to the 2001 annual U.S. report on human rights?. They developed an annotation scheme for low-level representation of opinions, and then proposed using
opinion-oriented scenario templates to act as a summary representation of the opinions. Possible ways of using the representations in multi-perspective QA are discussed. In related work,
C HAPTER 1. I NTRODUCTION
22
Stoyanov et al. [2005] analyzed characteristics of opinion questions and answers and show
that traditional FBQA techniques are not enough for multi-perspective QA. Results of some
initial experiments show that using filters that identify subjective sentences is helpful in multiperspective QA.
Summary
The typical work discussed here shows the state-of-the-art in NFQA. Most systems are investigating complex questions in specific domains or of particular types. Although interesting views
and approaches have been proposed, most work is at the initial stage, describing the general
framework or potential useful approaches to address characteristics of NFQA.
As mentioned in section 1.1, our work on NFQA is in the medical domain. Clinical QA
as an NFQA task, presents challenges similar to those of the tasks described in the previous
subsection. Our work is to investigate these challenges by addressing a key issue: what information is relevant? We do not attempt to elicit such information by deriving additional
questions, such as performing question decomposition [Hickl et al., 2004] or through interactive QA [Small et al., 2004]. Instead, we aim to identify the best information available in a
designated source to construct the answer to a given question. The next chapter will describe
our approach based on semantic class analysis.
1.6
Overview of contributions of thesis
This thesis focuses on a new branch of the question-answering task – NFQA. We show the
difference between NFQA and FBQA by analyzing new characteristics of NFQA. We claim
that answers to NFQA are usually more complex than named entities, and multiple pieces of
information are often needed to construct a complete answer. We propose a novel approach to
address these characteristics. Important subtasks in different modules of the new approach are
identified, and automatic methods are developed to solve the problems.
To achieve these goals, we propose to use semantic class analysis in NFQA and use frame
structure to represent semantic classes. We develop rule-based approaches to identify instances
of semantic classes in text. Two important properties of semantic classes (cores and polarity)
C HAPTER 1. I NTRODUCTION
23
are identified automatically. We show that the problem of relevance and redundancy in constructing answers is closely related to text summarization and build a summarization system to
extract important sentences.
The QA approach based on semantic class analysis An event or scenario describes relations of several roles. Therefore, roles and their relations represent the gist of a scenario. We
use semantic classes to refer to the essential roles in a scenario and propose an approach using semantic class analysis as the organizing principle to answer non-factoid questions. This
approach contains four major components:
• Detecting semantic classes in questions and answer sources
• Identifying properties of semantic classes
• Question-answer matching: exploring properties of semantic classes to find relevant
pieces of information
• Constructing answers by merging or synthesizing relevant information using relations
between semantic classes
We investigate NFQA in the context of clinical question answering, and focus on three semantic
classes that correspond to roles in the commonly accepted PICO format of describing clinical
scenarios. The three classes are: the problem of the patient, the intervention used to treat the
problem, and the clinical outcome. Interpretation of any treatment scenario can be derived
using the three classes. This semantic class-based approach is described in Chapter 2.
Extracting semantic classes and analyzing their relations
We use rule-based approaches
to identify clinical outcomes and relations between instances of interventions in sentences. In
QA, extracted clinical outcomes can be used directly to answer questions about outcomes of
interventions. In the combination approach of outcome identification that we developed, a set
of cue words that signal the occurrence of an outcome are collected and classified according to
their PoS tags. For each PoS category, the syntactic components it suggests are summarized to
derive rules for identifying boundaries of outcomes. This approach can potentially be applied to
C HAPTER 1. I NTRODUCTION
24
identify or extract any semantic class. We identify six common relationships between different
instances of interventions in a sentence and develop a cue-word based approach to identify
the relations automatically. These relationships will improve accuracy of matching between
questions and their answers. They can also improve document retrieval. After the index is built
for these relations, they can be queried directly. Instances of semantic classes and their relations
can be filled in predefined frame structures. Such information in free text is then represented by
a more-structured data format that is easier for further processing. The combination approach
and relation analysis are presented in Chapter 3.
Identifying cores of semantic classes We use the term core to refer to the smallest fragment of an instance of a semantic class that exhibits information rich enough for deriving a
reasonably accurate understanding of the class. We found that cores are an important property of semantic classes as they can be the only clues to find the right answers. In Chapter 4,
we show how cores of interventions, problems, and outcomes in a sentence can be identified
automatically by developing an approach exploring semi-supervised learning techniques. This
approach can be applied to identify cores of other semantic classes that have similar syntactic
constituents, and it can be adapted to other semantic classes that have different syntactic constituents. This approach can potentially be applied to other classification problems that aim
to group similar instances as well, e.g., word sense disambiguation. The concept of cores of
semantic classes is pertinent to many tasks in computational linguistics. For example, cores
are related to named entities. Some cores of semantic classes are named entities, while many
are not. Cores as a new type of semantic unit extends the idea of named entities and the applications that rely on named entity identification.
Detecting polarity of clinical outcomes A clinical outcome may be positive, negative or
neutral. Polarity is an inherent property of clinical outcomes. This information is mandatory
to answer questions about benefits and harms of an intervention. Information on negative
outcomes is often crucial in clinical decision making. We develop a method using a supervised
learning model to automatically detect polarity of clinical outcomes. We show that this method
has similar performance on different sources of medical text. We also identify a cause of the
C HAPTER 1. I NTRODUCTION
25
bottleneck of performance using supervised learning approaches in polarity classification. The
polarity detection task is discussed in Chapter 5.
Extracting components for answers We built explicit connection between text summarization and identifying answer components in NFQA, and construct a summarization system that
explores a supervised classification model to extract important sentences for answer construction. We investigate the role of clinical outcomes and their polarity in this task. The system is
presented in Chapter 6.
Chapter 2
Our approach for NFQA: semantic class
analysis
As discussed in Chapter 1, answers in NFQA are not named entities and often consist of multiple pieces of information. In response to these major characteristics of NFQA, we propose
to use frame-based semantic class analysis as the organizing principle to answer non-factual
questions.
We investigated NFQA in the context of clinical question answering. In this chapter, we
discuss the approach of semantic class analysis and how our work fits in the general QA framework.
2.1
Our approach of semantic class analysis
Clinical questions often describe scenarios. For example, they may describe relationships between clinical problems, treatments, and corresponding clinical outcomes, or they may be
about symptoms, hypothesized disease and diagnosis processes. To answer these questions,
essentially, we need an effective schema to understand scenario descriptions.
26
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
27
2.1.1 Representing scenarios using frames
Semantic roles Our principle in answering non-factual questions developed from the viewpoint that semantics of a scenario or an event is expressed by the semantic relationships between its participants, and such semantic relationships are defined by the role that each participant plays in the scenario. These relationships are referred to as semantic roles [Gildea and
Jurafsky, 2002], or conceptual roles [Riloff, 1999]. This viewpoint can date back to frame semantics, posed by Fillmore [1976] as part of the nature of language. Frame semantics provides
a schematic representation of events/scenarios that have various participants as roles. In our
work, we use frames as our representation schema for the semantic roles involved in questions
and answer sources.
Research on semantic roles has proposed different sets of roles ranging from the very
general to the very specific. The most general role set consists of only two roles: PROTOAGENT and PROTO-PATIENT [Dowty, 1991; Valin and Robert, 1993]. Roles can be more
domain-specific, such as perpetrators, victims, and physical targets in a terrorism domain. In
question-answering tasks, specific semantic roles can be more instructive in searching for relevant information, and thus more precise in pinpointing correct answers. Therefore, we take
domain-specific roles as our targets.
The treatment frame Patient-specific questions in EBM usually can be described by the socalled PICO format [Sackett et al., 2000] in the medical domain. In a treatment scenario, P
refers to the status of the patient (or the problem), I means an intervention, C is a comparison
intervention (if relevant), and O describes the clinical outcome. For example, in the following
question:
Q: In a patient with a suspected myocardial infarction does thrombolysis decrease the risk of
death?
the description of the patient is patient with a suspected myocardial infarction, the intervention
is thrombolysis, there is no comparison intervention in this question, and the clinical outcome
is decrease the risk of death. Originally, PICO format was developed for therapy questions
describing treatment scenarios and was later extended to other types of clinical questions such
as diagnosis, prognosis, and etiology. Representing clinical questions with PICO format is
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
28
widely believed to be the key to efficiently finding high-quality evidence [Richardson et al.,
1995; Ebell, 1999]. Empirical studies have shown that identifying PICO elements in clinical
scenarios improves the conceptual clarity of clinical problems [Cheng, 2004].
We found that PICO format highlights several important semantic roles in clinical scenarios, and can be easily represented using the frame structure. Therefore, we constructed a frame
based on it. Since C mainly indicates a comparison relation to I, we combined the comparisons
as one filler of the same slot intervention in the frame, connected by a specific relation. We focus on therapy-related questions and built a treatment frame that contains three slots, as shown
in Table 2.1.
Table 2.1: The treatment frame
P:
a description of the patient (or the problem)
I:
an intervention
O:
the clinical outcome
A slot in a frame designates a semantic class (corresponds to a semantic role or a conceptual
role), and relations between semantic classes in a scenario are implied by the design of the
frame structure. The treatment frame expresses a cause-effect relation: the intervention for the
problem results in the clinical outcome.
When applying this frame to a sentence, we extract constituents in the sentence to fill in
the slots in the frame. These constituents are instances of semantic classes. In this thesis, the
terms instances of semantic classes and slot fillers are used interchangeably. Some examples
of the instantiated treatment frame are as follows.
Sentence: One RCT [randomized clinical trial] found no evidence that low molecular weight heparin is superior to aspirin alone for the treatment of acute ischaemic stroke in people with atrial
fibrillation.
P: acute ischaemic stroke in people with atrial fibrillation
I: low molecular weight heparin vs. aspirin
O: no evidence that low molecular weight heparin is superior to aspirin
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
29
Sentence: Subgroup analysis in people with congestive heart failure found that diltiazem significantly increased death or reinfarction.
P: people with congestive heart failure
I: diltiazem
O: significantly increased death or reinfarction
Sentence: Thrombolysis reduces the risk of dependency, but increases the risk of death.
P: —
I: thrombolysis
O: reduces the risk of dependency, but increases the risk of death
The first example states the result of a clinical trial, while the second and third depict
clinical outcomes. We do not distinguish the two cases in this study, and treat them in the same
manner.
How is it related to information extraction (IE)? Our approach of semantic class analysis
has a close relation to IE, in which domain-specific semantic roles are often explored to identify
predefined types of information from text [Riloff, 1999]. Our approach shares the same view
with IE that semantic classes/roles are the keys to understand scenario descriptions. Frames are
also used in IE as the representation scheme. Nevertheless, in our work, as shown by the above
examples of treatment frames, the syntactic constituents of an instance of a semantic class can
be much more complex than those of traditional IE tasks, in which slot fillers are usually named
entities [Riloff, 1999; TREC, 2001]. Therefore, approaches based on such semantic classes go
beyond named-entity identification, and thus will better adapt to NFQA. In addition, extracting
instances of semantic classes from text is not the ultimate goal of QA. Frame representation of
semantic classes provides a platform for matching between questions and answers in our QA
system. We propose to conduct further analysis on semantic classes to search for answers to
non-factual questions, which will be described in the following subsection.
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
30
2.1.2 Main components of a QA system guided by semantic class analysis
We propose to use semantic class analysis to guide the process of searching for answers to
non-factual questions.
With semantic class analysis as the organizing principle, we identify four main components
of our QA system:
• Detecting semantic classes in questions and answer sources
• Identifying properties of semantic classes
• Question-answer matching: exploring properties of semantic classes to find relevant
pieces of information
• Constructing answers by merging or synthesizing relevant information using relations
between semantic classes
To search for the answer to a question, the question and the text in which the answer may
occur will be processed to detect the semantic classes. A semantic class can have various
properties. These properties can be extremely valuable in finding answers, which we will
discuss in detail in Chapter 4, 5, and 6. In the matching process, the question scenario will be
compared to an answer candidate, and pieces of relevant information should be identified by
exploring properties of the semantic classes. To construct the answer, relevant information that
has been found in the matching process will be merged or synthesized to generate an accurate
and concise answer. The process of synthesizing scenarios relies on comparing instances of
semantic classes in these scenarios. For example, two instances are exactly the same or one is
the hypernym of the other.
Scenario questions are common in other domains as well. For instance, questions about
shipping events often depict relations between provider, receiver, and means; questions on
events like criticizing often contain a reviewer, an object, the reason, and the manner. Frame
semantics is a general representation schema for scenarios. Therefore, we expect that the main
components in our QA approach can be applied to scenario questions in other domains rather
easily.
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
31
2.1.3 The EPoCare Project
Our work is part of the EPoCare project (“Evidence at the Point of Care”) at the University of Toronto. The project aims to provide clinicians fast access at the point of care to the
best available medical information in published literature. Clinicians will be able to query
sources that appraise the evidence about the treatment, diagnosis, prognosis, etiology, and
prevalence of medical conditions. In order to make the system available at the point of care,
the question-answering system will be accessible using hand-held computers. The project is an
interdisciplinary collaboration that involves research in several disciplines. Project members
in Industrial Engineering and Cognitive Psychology are investigating the design of the system
through a user-centered design process, in which requirements are elicited from end users who
are also involved in the evaluation of the prototypes. Project members in Knowledge Management and Natural Language Processing aim to ensure that the answers to queries are accurate
and complete. And project members in Health Informatics will test the influence of the system
on clinical decision-making and clinical outcomes.
Figure 2.1 shows the architecture of the system. There are three main components in the
system. The data sources are stored in an XML document database. The EPoCare server
uses this database to provide answers to queries posed by clinicians. The knowledge base is
the source of medical terminologies.
Data sources The current data sources include the reviews of experimental results for clinical problems that are published in Clinical Evidence (CE) (version 7) [Barton, 2002], and
Evidence-based On Call (EBOC) [Ball and Phillips, 2001].
• CE is a publication that reviews the current state of knowledge about the prevention
and treatment of clinical conditions. It is a source of evidence on the effects of clinical
interventions and it is updated every six months. The main content of CE is described
in natural language. Evidence in CE is organized by a hierarchy structure of disease
categories. In this structure, specific diseases are grouped together under each general category of disease, as shown in figure 2.2. For each specific disease, the effects
of various interventions are summarized. CE is the text source that is used in most
experiments reported in this thesis.
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
32
Client Application
clinical query
clinical answers
EPoCare Server
Retriever
expansion of keywords
candidate
documents
Query−Answer
Matcher
candidate answers
Answer Generator
ToX query / answer
ToX Engine
UMLS
CE
EBOC
Knowledge base
Data Sources
Figure 2.1: EPoCare system architecture.
• EBOC is another source that supports EBM. It provides the best available evidence
on important topics in clinical practice by reviewing and summarizing knowledge in
several databases, including the ‘Best Evidence’ CD-ROM, the Cochrane Library, and
PubMed. Topics in EBOC are arranged alphabetically, indexed by disease area. Unlike
CE, which has a focus on treatments, EBOC covers prevalence, clinical features, investigations, therapy, prevention, and prognosis. Summaries of the evidence are written
in natural language, and are often accompanied by tables containing data derived from
the original studies.
Both data sources are stored with XML mark-up in the database. The XML database is
manipulated by ToX, a repository manager for XML data [Barbosa et al., 2001]. Repositories
of distributed XML documents may be stored in a file system, a relational database, or remotely
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
33
Acute atrial fibrilation
Cardiovascular disorders
Acute myocardial infarction
Stroke management
Heart failure
...
Acute gastroenteritis in children
Child health
Acute otitis media
Asthma in children
...
Acute appendicitis
Digestive system disorders
Anal fissure
Colonic diverticular disease
...
...
Figure 2.2: Disease categories in Clinical Evidence
on the Web. ToX supports document registration, collection management, storage and indexing
choice, and queries on document content and structure.
EPoCare server In the EPoCare server, the Knowledge Management team takes care of
keyword-based searching. A clinical query from the client is processed to form a database
query of keywords. The query is sent by the retriever to the XML document database to
retrieve relevant documents (e.g., a complete or partial section in CE) in the data sources using
keyword matching. The results are then passed to the query–answer matcher to find the answer
candidates. Finally, the best answer is determined and returned to the user.
The role of natural language processing is to allow the system to accept queries expressed in
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
34
natural language and to better identify answers in its natural-language data sources. After relevant documents are retrieved using the keyword-based matching, sentences in these documents
will be processed using natural language processing techniques to find accurate and concise
answers. Our work described in the following chapters can be adapted to several modules of
the EPoCare system, including the query-answer matcher and the answer extractor.
Knowledge base The Unified Medical Language System (UMLS) is a knowledge base of
medical terminologies. It is the major knowledge base in our work. UMLS contains three
knowledge sources.
• The Metathesaurus is the central vocabulary component that contains information about
biomedical and health-related concepts and the relationships among them. More than
one name can be used to refer to the same concept. Metathesaurus links them together.
There are 11 types of relationships between concepts in Metathesaurus, including synonymy, broader, and narrower. Each concept in the Metathesaurus is assigned to at
least one semantic type from another component of UMLS – the Semantic Network.
• The Semantic Network is a network of the general categories or semantic types, such
as mental disability and pathological functions, to which all concepts in the Metathesaurus have been assigned. It provides a consistent categorization of all concepts represented in the UMLS Metathesaurus and the important relationships between them. The
2003AA release of the Semantic Network contains 135 categories and 54 relations. In
the Network, the categories are the nodes, and the relationships between them are the
links. The primary link in the Network is the isa link. In addition, non-hierarchical
relations are also identified, which belong to five major categories: physically related
to, spatially related to, temporally related to, functionally related to, and conceptually
related to.
• The SPECIALIST lexicon contains syntactic information about biomedical terms. It
covers commonly occurring English words and biomedical vocabulary. The lexicon
entry for each word or term records the syntactic, morphological, and orthographic
information.
C HAPTER 2. O UR APPROACH FOR NFQA: SEMANTIC CLASS ANALYSIS
35
The following chapters discuss our work in three of the main components of our QA system.
Figure 2.3 shows how this work fits in the general QA architecture.
Question
Set
Identifying
Semantic Classes
(Chapter 3)
Question
Processing
Q-A
Matching
Extracting Cores
(Chapter 4)
Detecting Polarity
(Chapter 5)
Document
Set
Document
Processing
Answer
Generation
Extracting
Answer
Components
(Chapter 6)
Figure 2.3: Our work in the QA framework
Chapter 3
Identifying semantic classes in text:
filling the frame slots
This chapter discusses two problems in filling the treatment frame: identifying semantic classes
in text and analyzing relations between instances of a semantic class. In semantic class identification, we focus on clinical outcomes, as outcomes are often expressed by more complex
syntactic structures and are more difficult to label. In medical text, more than one intervention
is often mentioned in the treatment of a disease, and various types of relations are involved
between the interventions. These relations are analyzed automatically. We use rule-based approaches in these tasks.
3.1
Identifying clinical outcomes using
a combination approach
In medical text, the appearance of some words is found often to be a signal of the occurrence
of an outcome, and usually several words signal the occurrence of one single outcome. The
combination approach that we applied for identifying outcomes is based on this observation.
Our approach does not extract the whole outcome at once. Instead, it tries to identify the
different parts of an outcome that may be scattered in the sentence, and then combines them to
form the complete outcome.
36
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 37
In the combination approach, different pieces of an outcome are identified by some lexical identifiers, which are referred to as cue words. Each occurrence of a cue word suggests
a portion of the expression of the outcome. Detecting all of them will increase the chance of
obtaining the complete outcome. Also, different occurrences of cue words provide more evidence of the existence of an outcome. We evaluate the two phases of outcome identification
separately. The first step is detecting the occurrence of outcomes, and the second is determining
the boundaries of outcomes.
In the experiment, the text we use is from Clinical Evidence (CE). Two sections of CE
were analyzed for detection of outcome. Outcome information in the text was annotated by a
clinician. About two-thirds of each section (267 sentences in total) was taken as the analysis
examples to construct the rules, and the rest (156 sentences) as the test set.
3.1.1 Detecting clinical outcomes in text
Collecting cue words We manually analyzed the analysis examples, and found that cue
words of clinical outcomes belong to three PoS categories: noun, verb, and adjective. The
cue words we found in the analysis are listed in Figure 3.1. All the inflectional variants of the
cues are used as identifiers in the experiment.
Nouns: death benefit dependency outcome evidence harm difference risk deterioration
mortality disability independence survival significance
Verbs: improve reduce prevent produce increase decrease affect
Adjectives: beneficial harmful negative adverse superior effective
Figure 3.1: Cue words for detecting clinical outcomes
In the following examples, cue words are highlighted.
(1) Thrombolysis reduces the risk of dependency, but increases the risk of death.
(2) Lubeluzole has also been noted to have adverse outcome, especially at higher doses.
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 38
Table 3.1: Results of identifying outcomes in CE
False
False
Method
Correct
Positives
Negatives
Precision%
Recall%
F-score%
Accuracy%
baseline
81
75
0
52 (81/156)
100
68
52
combination approach
67
14
14
83 (67/81)
83
83
82
(3) Several small comparative RCTs [randomized clinical trials] have found sodiumcromoglicate to be less effective than inhaled corticosteroids in improving symptoms and
lung function.
(4) In the systematic review of calcium channel antagonists, indirect and limited comparisons of intravenous versus oral administration found no significant difference in
adverse events.
The last two examples are different from other examples in that they express the outcomes
of clinical trials, which we refer to as “results” in the following description when necessary. A
“result” might contain a clinical outcome within it, as results often involve a comparison of the
effects of two (or more) interventions on a disease.
Evaluating the outcome detection task We evaluated the cue word method of detecting the
outcome on the test set. The result is shown in Table 3.1. A sentence that contains a clinical
outcome is a positive case. Eighty-one sentences in the test set contain outcomes, which is
52% of all the test sentences. This was taken as the baseline of the evaluation: assigning all
sentences in the test set to positive. By contrast, the accuracy of the cue word approach is 82%.
In error analysis, we found two main reasons that some outcomes were missed in the identification. One is that some outcomes do not have any cue word:
(5) Gastrointestinal symptoms and headaches have been reported with both montelukast
and zafirlukast.
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 39
This example describes that two adverse events are associated with the treatment, which implies a negative clinical outcome. This outcome is only expressed implicitly, and therefore
missed by the cue word-based approach.
The other reason is that although some words might be regarded as cue words, we did not
include them in our set; for example, fewer and higher. Adjectives were found to have the most
irregular usages in identifying outcomes. It is common for them to modify both interventions
and outcomes, as shown in the following examples:
(6) Growth was significantly slower in children receiving higher dose inhaled corticosteroids.
(7) At 12 weeks, mean morning PEFR (peak expiratory flow rate) was 4% higher in the
salmeterol group.
The word higher only signals a clinical outcome in the second example. Other adjectives such
as less, more, lower, shorter and longer have similar problems. If they are taken as identifiers
of outcomes then some false positives are very likely to be generated. However, if they are
excluded, some true outcomes will be missed.
There were 14 false positives in the result of the experiment. The main cause was that some
sentences contain cue words, and yet they did not provide any useful information:
(8) We found that the balance between benefits and harms has not been clearly established
for the evacuation of supratentorial haematomas.
(9) The third systematic review did not evaluate these adverse outcomes.
As mentioned at the beginning of this subsection, currently, the cue words for detecting
clinical outcomes were collected manually. In the next step of our work, we will investigate
automatic approaches such as bootstrapping to find cues.
3.1.2 Determining the textual boundary of clinical outcomes
After the occurrence of clinical outcomes is detected using the cue words, the next problem
is to determine their textual boundaries. Again, we rely on cue words to find the clue. As
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 40
mentioned before, a single clinical outcome often has several cue words in its expressions. The
idea is that if the fragment of clinical outcome suggested by each cue word is identified, then
the complete outcome can be obtained by combining or merging these fragments. Because the
cue words belong to three PoS groups, we investigate the syntactic structure of the fragments
in a clinical outcome that may be suggested by each of the groups.
Developing boundary detection rules In the analysis set, we obtained the syntactic structure
of each sentence, as well as the PoS tags of words in the sentence using the Apple Pie parser
[Sekine, 1997]. This parser produces phrase and PoS information that is needed for our work.
It has various output formats, and one of them (shown in the following example) fits our task
well and is easy to process. Some examples of the output of the parser with PoS and phrase
information are listed in Figure 3.2. In the parsing trees, syntactic tags are all capitalized while
Sentence:
Thrombolysis reduces the risk of dependency, but increases the risk of death.
Output of the parser:
(S (NPL (NNPX Thrombolysis)) (VP (VP (VBZ reduces) (NP (NPL (DT the) (NN risk)) (PP
(IN of) (NPL (NN dependency))))) (, -COMMA-) (CC but) (VP (VBZ increases) (NP (NPL
(DT the) (NN risk)) (PP (IN of) (NPL (NN death)))))) (. -PERIOD-))
Sentence:
Lubeluzole has also been noted to have adverse outcome, especially at higher doses.
Output of the parser:
(S (NPL (NNPX Lubeluzole)) (VP (VBZ has) (ADVP (RB also)) (VP (VBN been) (VP (VBN
noted) (TOINF (VP (TO to) (VP (VB have) (NPL (JJ adverse) (NN outcome)) (, -COMMA-)
(PP (ADVP (RB especially)) (IN at) (NPL (JJR higher) (NNS doses))))))))) (. -PERIOD-))
Figure 3.2: Examples of output of Apple Pie Parser
words are not. Each word is attached with a PoS tag, which immediately precedes the word.
Phrases are marked by higher level parentheses starting with phrase tags. Syntactic tags in the
figure are explained in Appendix B.
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 41
Three general rules are derived to guide the boundary determination from the parsing trees
of all sentences in the analysis example set:
• If a cue is a noun: The noun phrase that contains the noun will be part of the outcome.
• If a cue is a verb: The verb and its object (if the verb is in active voice) or its subject
(if the verb is in passive voice) together constitute one portion of the outcome.
• If a cue is an adjective: The corresponding adjective phrase or the noun phrase belongs
to the outcome.
In the first sentence in Figure 3.2, the verb reduces is a cue. According to the rule for
verbs, the noun phrase that immediately follows it is part of the outcome. Therefore, the risk of
dependency is included in the outcome. Similarly, the cue word increases identifies the risk of
death as part of the outcome. By combining the two parts, the complete description of outcome
is detected.
Cue words for the results of clinical trials are processed in a slightly different way. For
example, for difference and superior, any immediately following prepositional phrase is also
included in the results of the trial, as shown in Figure 3.3. The algorithm of the combination
approach is described in Appendix D.
Sentence:
In the systematic review of calcium channel antagonists, indirect and limited comparisons of
intravenous versus oral administration found no significant difference in adverse events.
Output of the parser:
(S (PP (IN In) (NP (NPL (DT the) (JJ systematic) (NN review)) (PP (IN of) (NPL (NN
calcium) (NN channel) (NNS antagonists))))) (, -COMMA-) (NP (NPL (ADJP (JJ indirect)
(CC and) (JJ limited)) (NNS comparisons)) (PP (IN of) (NPL (ADJP (JJ intravenous) (CC
versus) (JJ oral)) (NN administration)))) (VP (VBD found) (NPL (DT no) (JJ significant) (NN difference)) (PP (IN in) (NPL (JJ adverse) (NNS events)))) (. -PERIOD-))
Figure 3.3: An example of cue word difference
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 42
Table 3.2: Results of boundary detection of correctly identified outcomes in CE
Type of Overlap
Number
Percentage
Exact match
26
39
A entirely within B
19
28
B entirely within A
13
19
Both partially within the other
8
12
No match
1
1
A: Identified fragments; B: true boundary
Evaluating the boundary detection task Table 3.2 shows the result of boundary detection
for those outcomes that were correctly identified (i.e., the true positives in the previous outcome
detection task). The true boundary is the boundary of an outcome that was annotated manually.
The no match case means that there is a true outcome in the sentence but the program missed the
correct portions of text and marked some other portions as the outcome. The program identified
39% of the boundaries exactly the same as the true boundaries. In 19% of the samples, the true
boundaries were entirely within the identified fragments. The spurious text in them (the text
that was not in the true boundary) was found to be small in many cases, both in terms of
number of words and in terms of the importance of the content. The average number of words
correctly identified was 7 for each outcome and the number of spurious words was 3.4. The
most frequent content in the spurious text was the intervention used to obtain the outcome. In
the following examples, text in “hi” is the outcome (result) identified automatically, and text in
“{}” is spurious.
(10) The RCTs found hno significant adverse effects {associated with salmeterol}i.
(11) The second RCT also found hno significant difference in mortality at 12 weeks {with
lubeluzole versus placebo}i.
In the boundary detection task, again, adjectives are most problematic because of the great
variation in the expression of outcomes they suggest. In the following examples, the true
boundaries of outcomes are indicated by “[ ]”, and adjectives are highlighted.
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 43
(12) Small RCTs with physiological rather than clinical end points found that giving β2
agonists by metered dose inhaler with spacer to wheezy infants was [h effective i],
[with less likelihood] than nebulisation [to show h transient reduction of lung function
i].
(13) Nebulised β2 agonists are known to cause [tachycardia, tremor, and hypokalaemia],
but [ h serious adverse effects i are rare].
In sentence (12), the adjective effective is part of the outcome. In sentence (13), the clause
that contains the adjective adverse is part of the outcome.
The correctness of the output of the parser also affects the performance, as shown in the
following example:
Sentence:
RCTs found no evidence that lubeluzole improved clinical outcomes in people with acute ischaemic
stroke.
Output of the parser:
(S (NPL (NNPX RCTs)) (VP (VBD found) (NPL (DT no) (NN evidence)) (NPL (DT that) (JJ
lubeluzole) (JJ improved) (JJ clinical) (NNS outcomes)) (PP (IN in) (NP (NPL (NNS people)) (PP
(IN with) (NPL (JJ acute) (JJ ischaemic) (NN stroke)))))) (. -PERIOD-))
The verb improve was incorrectly assigned to be an adjective in a noun phrase. Thus improve as
a verb cue word was missed in identifying the outcome. However, another cue word outcomes
was found, so the whole noun phrase containing outcomes was identified as the outcome. This
example also shows that missing one cue word in identifying the outcome can be corrected by
the occurrence of other cue words in the combination approach.
Related work
Machine-learning approaches and rule-based methods have been used for similar problems.
Gildea and Jurafsky [2002] used a supervised learning method to learn both the identifier of
the semantic roles defined in FrameNet such as theme, target, goal, and the boundaries of
the roles [Baker et al., 2003]. A set of features were learned from a large training set, and
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 44
then applied to the unseen data to detect the roles. The performance of the system was quite
good. However, it requires a large training set for related roles. It is usually expensive and
time-consuming to obtain a large manually annotated data set.
Rule-based methods are explored in information extraction (IE) to identify roles to fill in
slots in some pre-defined templates [Català et al., 2003]. The rules are represented by a set
of patterns, and template role identification is usually conducted by pattern matching. Slots
indicating roles are embedded in these patterns. Text that satisfies the constraints of a pattern
will be identified, and the contents corresponding to the slots are extracted. This approach
has been proved to be effective in many IE tasks. However, pattern construction is very timeconsuming. In order to extract the roles and only the roles from text, their expressions have
to be customized specifically in patterns. Targets consisting of complex syntactic constituents,
e.g., clinical outcomes, will result in increasing difficulties in pattern construction, and less
coverage of the patterns.
In our combination approach, instead of building one pattern to extract complete information of a target, as was done in most IE systems, we constructed simpler rules to identify
portions of the target and then combine them to get the complete information. We expect this
strategy to release some burden of manually creating patterns, especially for tasks having complex targets. In addition, since it is a rule-based approach, it does not need a large manually
annotated training set. A limitation of this approach is that some connections between different portions of an outcome may be missing. Also, a different set of cue words may need to be
collected when adapting to a new domain.
3.2
Analysis of Relations
More than one instance of semantic classes often occurs in a sentence; some instances are of
the same class, while some of them are not. For those of different semantic classes, i.e., intervention, disease, and outcome, they often follow the relation implied by the frame structure.
In our treatment frame, it is a cause-effect relation: the use of the intervention to the disease
results in the outcome. For those of the same semantic class, e.g., intervention, we found that
various relations occur. As discussed in Section 2.1, scenarios are about semantic classes and
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 45
their relations. In our approach of semantic class analysis, understanding such relations is an
important part of interpreting scenarios. These relations are the target of this section. We
only evaluate relations between different instances of intervention in a sentence, as we observe
that a sentence often mentions more than one intervention. Relations between diseases can be
analyzed in a similar way although they occur much less often than interventions.
Text from CE was analyzed manually to understand what relations are often involved and
how they are represented. Then, an approach was developed to automatically identify the
relations. The text for the analysis and test is the same as in the outcome identification task.
Interventions in the text were annotated by a clinician.
Collecting cue words and symbols As with outcome identification, we found that these relations can be identified by a group of cue words or symbols. For example, the word plus refers
to the COMBINATION of two or more interventions, the word or, as well as a comma, often suggests the ALTERNATIVE relation, and the word versus (or v) usually implies a COMPARISON
relation, as shown in the following examples:
(14) The combination of aspirin plus streptokinase significantly increased mortality at 3
months.
(15) RCTs found no evidence that calcium channel antagonists, lubeluzole, aminobutyric
acid agonists, glycine antagonists, or N-methyl-D-aspartate antagonists improve clinical outcomes in people with acute ischaemic stroke.
(16) One systematic review found no short or long term improvement in acute ischaemic
stroke with immediate systemic anticoagulants (unfractionated heparin, low molecular
weight heparin, heparinoids, or specific thrombin inhibitors) versus usual care without
systemic anticoagulants.
It is worth noting that in CE, the experimental conditions are often explained in the description
of the outcomes, for example:
(17) Growth was significantly slower in children receiving higher dose inhaled corticosteroids (3.6cm, 95% CI 3.0 to 4.2 with double dose beclometasone v 5.1cm, 95% CI
4.5 to 5.7 with salmeterol v 4.5cm, 95% CI 3.8 to 5.2 with placebo).
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 46
(18) It found that the addition for 4 weeks of oral theophylline versus placebo increased
the mean number of symptom free days (63% with theophylline v 42% with placebo;
P=0.02).
(19) Studies of adults with poor control on low dose inhaled steroid (see salmeterol v high
dose inhaled corticosteroids under adult asthma) have found greater benefit with additional long-acting β2 agonists than with higher doses of inhaled steroid.
These conditions are usually in parentheses. They are often phrases and even just fragments
of strings that are not represented in a uniform way with the other parts of the sentence. Their
behavior is more difficult to capture and therefore the relations among the concepts in these descriptions are more difficult to identify. Because they usually are examples and data, omission
of them will not affect the understanding of the whole sentence in most cases.
Six common relations and their cue words were found in the text which are shown in Table
3.3. Cue words and symbols between interventions were first collected from the training text.
Then the relations they signal were analyzed. Some cue words are ambiguous, for example,
and, and with. It is interesting to find that and in the text when it connects two interventions
usually suggests an alternative relation rather than a combination relation, as in the example:
(20) Both salmeterol and beclometasone improved FEV1 compared with placebo, but the
difference between beclometasone and salmeterol was not significant.
Compared with versus, plus, etc., and and with are weak cues as many of their appearances
in the text do not suggest a relation between two interventions.
Experiment On the basis of this analysis, an automatic relation analysis process was applied
to the test set. The test set is the same as in outcome identification. In the experiment, if a
cue presents between two interventions in a sentence, the relation of the interventions will be
detected. To deal with the case that more than one cue appear between two interventions, we
assigned priorities to cue words/symbols according to how strong they are. A cue with higher
priority determines the relation. And and with get lower priority compared to other cues. And
has higher priority than with. For “,” and “(”, they are cues only when they are the only symbols
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 47
Table 3.3: Cue words/symbols for relations between interventions
Relation(s)
Cue Words/Symbols
COMPARISON
superior to, more than, versus,
compare with, between . . . and . . .
ALTERNATIVE
or, “,”, and
COMBINATION
plus, add to, addition of . . . to . . . ,
combined use of, and, with, “(”
SPECIFICATION
with, “(”
SUBSTITUTE
substitute, substituted for
PREFERENCE
rather than
between two interventions. Therefore they do not need to be assigned priorities. Other cues
are not assigned priorities since they usually do not co-occur between two interventions. For
ambiguous cues and, with, and “(”, we assign the most frequent relation they indicate in the
analysis examples to any occurrence of them in the test set. Therefore, and suggests alternative
relations, with and “(” indicate specification relations.
The test process was divided into two parts: one took parenthetical descriptions into account (case 1) and the other one did not (case 2). In the evaluation, for sentences that contain
at least two interventions, “correct” means that the relation identified automatically is the same
as marked by the annotator, “wrong” indicates that the two are different. In a “missing” case,
a relation is ignored by the automatic approach. We did not evaluate the relation between any
two interventions in a sentence; instead, we only considered two interventions that are related
to each other by a cue word or symbol1 (including those connected by cue words other than the
set collected from the training text). The results of the two cases are shown in Table 3.4. Most
errors are because of the weak indicators with and and. As in the outcome identification task,
both the training and test sets are rather small, as no standard annotated text is available.
Some of the surface relationships in Table 3.3 reflect deeper relationships of the semantic
1
There is only one implicit relation (a relation without a cue word identifier) for case 1 and case 2 respectively.
C HAPTER 3. I DENTIFYING SEMANTIC CLASSES IN TEXT: FILLING THE FRAME SLOTS 48
Table 3.4: Results of relation analysis
Correct
Wrong
Missing
False Positive
case 1
49
7
10
9
case 2
48
7
3
6
classes. For example, COMPARISON, ALTERNATIVE, and PREFERENCE imply that the two
(or more) interventions have some common effects on the disease(s) that are treated. The
SPECIFICATION
relation, on the other hand, suggests a hierarchical relation between the first
intervention and the following ones, in which the first intervention is a higher-level concept and
the following interventions are at a lower level. For example, in example (16), systemic anticoagulants is a higher-level concept, unfractionated heparin, low molecular weight heparin, etc.,
are examples of it that lie at a lower level.
3.3
Summary
This chapter describes our work in identifying clinical outcomes and analyzing relations between interventions. In question-answering, this information will be extracted to fill in the
question frame and frames of potential answers. In addition, the relations can be indexed to
improve document retrieval by supporting direct relation search. For example, if a user is
interested in a comparison study of two interventions, then specifying both the relation and
the interventions as the searching strategy will get more accurate results than just looking for
the interventions. It can be very important for medical information retrieval, as such relations
occur frequently in the text.
Chapter 4
Cores of semantic classes
In this chapter, we discuss a property of semantic classes – their cores.
In a frame structure, the slots in question and answer frames can be filled with either complete or partial information. Consider the following example, where parentheses delimit each
instance of a semantic class (a slot filler) and the labels P (problem description), I (an intervention), O (the clinical outcome) indicate the type of the instance:
Sentence:
Two systematic reviews in (people with AMI)P investigating the use of (calcium channel blockers)I
found a (non-significant increase in mortality of about 4% and 6%)O.
Complete slot fillers:
P: people with AMI
I: calcium channel blockers
O: a non-significant increase in mortality of about 4% and 6%
Partial slot fillers:
P: AMI
I: calcium channel blockers
O: mortality
The partial slot fillers in this example contain the smallest fragments of the corresponding
complete slot fillers that exhibit information rich enough for deriving a reasonably precise
answer. We use the term core to refer to such a fraction of a slot filler (instance of a semantic
49
C HAPTER 4. C ORES OF SEMANTIC CLASSES
50
class).
4.1
Importance of cores
As discussed in Chapter 1, before the matching process, keyword-based document retrieval is
usually performed to find relevant documents that may contain the answer to a given question.
Keywords in the retrieval are derived from the question. Cores of semantic classes can be
extremely valuable in searching for such documents for complex question scenarios, as shown
in the following example.1
Question scenario:
A physician sees a 7-year-old child with asthma in her office. She is on flovent and ventolin
currently and was recently discharged from hospital following her fourth admission for asthma
exacerbation. During the most recent admission, the dose of flovent was increased. Her mother
is concerned about the impact of the additional dose of steroids on her daughter’s growth. This
is the question to which the physician wants to find the answer.
For a complex scenario description like this, the answer could be missed or drowned in irrelevant documents found by inappropriate keywords derived from the question. However, the
search can be much more effective if we have the information of cores of semantic classes, for
example, P: asthma, I: steroids, O: growth.
Similarly, semantics presented in cores can help filter out irrelevant information that cannot
be identified by searching methods based on simple string overlaps.
(21) In patients with myocardial infarction, do β blockers reduce all cause mortality and
recurrent myocardial infarction without adverse effects?
(22) In someone with hypertension and high cholesterol, what management options will
decrease his risk of stroke and cardiac events?
In question (21), the first occurrence of myocardial infarction is a disease and the second is part
of the clinical outcome. In question (22), stroke is part of the clinical outcome rather than a
disease to be treated, as it usually is. Obviously, string matching cannot distinguish between the
1
The scenario is an example used in usability testing in the EPoCare project at the University of Toronto.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
51
two cases. By identifying and classifying cores of semantic classes, the relations between these
important semantic units in the scenarios are very clear. Therefore, documents or passages that
do not contain myocardial infarction or stroke as clinical outcomes can be discarded.
In addition, identifying cores of semantic classes in documents can facilitate the questionanswer matching process. Some evidence relevant to the above question scenario on asthma is
listed below, where boldface indicates a core:
Evidence1: A more recent systematic review (search date 1999) found three RCTs comparing the effects of becolmetasone and non-steroidal medication on linear growth in children
with asthma (200 µg twice daily, duration up to maximum 54 weeks) suggesting a short-term
decrease in linear growth of -1.54 cm a year.
Evidence2: Two systematic reviews of studies with long term follow up and a subsequent longterm RCT have found no evidence of growth retardation in asthmatic children treated with
inhaled steroids.
The evidence sentences here are from CE [Barton, 2002]. The clinical outcomes mentioned in
the evidence have very different phrasings — yet both pieces of the evidence are relevant to the
question. The pieces of evidence describe two distinct outcomes — that short-term decrease in
growth is found and that there is no effect on growth in some long-term studies. Missing any of
the outcomes will lead to an incomplete answer for the physician. Here, cores of the semantic
classes provide the only clue that both pieces of this evidence are relevant to this question and
should be included in the answer. Hence, a complete description of semantic classes does not
have to be found. In fact, such a description with more information could make the matching
harder to find because of the different expressions of the outcomes.
Finally, cores of semantic classes in a scenario are connected to each other by the relations
embedded in the frame structure. The frame of the treatment scenario contains a cause-effect
relation: an intervention used to treat a problem results in a clinical outcome.
In this chapter, we propose a method to automatically identify and classify the cores of
semantic classes according to their context in a sentence. We take the treatment frame as
an example, in which the goal is to identify cores of interventions, problems, and clinical
outcomes. For ease of description, we will use the terms intervention-core, disease-core, and
outcome-core to refer to the corresponding cores. We work at the sentence level, i.e., we
C HAPTER 4. C ORES OF SEMANTIC CLASSES
52
identify cores in a sentence rather than a clause or paragraph. Two principles are followed
in developing the method. First, complete slot fillers do not have to be extracted before core
identification. Second, we aim to reduce the need for expensive manual annotation of training
data by using a semi-supervised approach.
4.2
Architecture of the method
In our approach, we first collect candidates of the target cores from sentences under consideration. For each candidate, we classify it as one of the four classes: intervention-core, diseasecore, outcome-core, or other. In the classification, a candidate will get a class label according
to its context, its UMLS semantic types, and the syntactic relations in which it participates.
Figure 4.1 shows the architecture of the approach.
Preprocessing
Classification
PoS Tagging
Syntactic Relations
words
cores
candidates
p(c | n) or tf · idf
Context Features
(noun phrases)
UMLS
Domain Features
Figure 4.1: Architecture of the approach of core identification
4.3
Preprocessing
In the preprocessing, all words in the data set are examined. The first two steps are to reduce
noise, in which some of the words that are unlikely to be part of real cores are filtered out.
Then, the rest are mapped to their corresponding concepts, and these concepts are candidates
of target cores.
PoS tagging Our observation is that cores of the three types of slot fillers are usually nouns
or noun phrases. Therefore, words that are not nouns are first removed from the candidate set.
PoS tags are obtained by using Brill’s tagger [Brill, 1993].
C HAPTER 4. C ORES OF SEMANTIC CLASSES
53
Filtering out some bad nouns This step is the second attempt to remove noise from the
candidate set. Nouns that are unlikely to be part of real cores are considered as bad candidates.
Two research options of measures are used to evaluate how good a noun is.
• Extended tf · idf .
Let DocSet = {D1 , · · · , Dd } be a set of documents, and N ounSet = {N1 , · · · , Nn }
be the set of nouns in DocSet. For any i ∈ [1..n] and j ∈ [1..d], let o #
ij be the number
of occurrences of Ni in Dj , and c#
i be the number of documents where Ni appears.
The score of Ni is defined as
scorei = maxj∈[1..d] (1 + tf ij ) × idf i
where
tf ij

 log o#
ij
=
 0
: o#
ij > 0
:
o#
ij
=0
and idf i = log(d/c#
i ).
The formula (1 + tf ij ) × idf i is taken from [Manning and Sch´’utze, 1999], and it is the
traditional measure of informativeness of a word with regard to a document. After this
value is calculated for a noun in each document, the highest value of all the documents
is taken as the final score of the noun. Nouns with scores lower than a threshold are
removed from the candidate set. The threshold was set manually after observing the
scores of some nouns that frequently occur in the text. CE text is used to get the score
of a noun. For this, 47 sections in CE are segmented to 143 files of about the same
size. Each file is treated as a document. This measure is referred to as tf · idf in later
description.
• Domain specificity. We calculate the conditional probability p(c | n)= p(c , n)/p(n),
where c is the medical class, and n is a noun. It is the probability that a document is
in the medical domain c given it contains the noun n. Intuitively, intervention-cores,
disease-cores, and outcome-cores are domain-specific, i.e., a document that contains
them is very likely to be in the medical domain. For example, morbidity, mortality, aspirin, and myocardial infarction are very likely to occur in a medicine-related context.
This measure intends to keep highly medical domain-specific nouns in the candidate
C HAPTER 4. C ORES OF SEMANTIC CLASSES
54
set. A noun is a better candidate if the corresponding probability is high. Text from
two domains is needed in this measure: medical text, and non-medical text. In our experiment, we use the same 47 sections in CE as the medical class text (separated into
143 files of about the same size). For the non-medical class, we use Reuters-21578
text collection. The collection has 21,578 documents. These documents appeared on
the Reuters newswire in 1987, and they are collected 1000 to a file. We use it because
newswire stories are mainly in the general domain. One file (1000 documents) in the
Reuters collection are randomly selected for the calculation. Nouns whose probability
values are below a threshold (determined in the same manner as in the tf · idf measure)
are filtered out.
Mapping to concepts To this point, the candidate set consists of nouns. In many cases,
nouns are part of noun phrases (concepts) that are better candidates of cores. For example, the
phrase myocardial infarction is a better candidate of a disease-core than the noun infarction.
Therefore, a noun is mapped to its corresponding UMLS concept in the sentence. All the
concepts form the candidates of cores to be classified.
To find the concepts, a sentence is processed by the software MetaMap [Aronson, 2001].
MetaMap maps biomedical text to concepts in the UMLS Metathesaurus and finds their semantic types in the semantic network of UMLS. The major steps in the mapping conducted by
MetaMap is outlined as follows.
1. Parsing. Text is parsed to get (mainly) simple noun phrases using the SPECIALIST
minimal commitment parser [National Library of Medicine, 2004].
2. Variant generation. For each phrase, variants are generated using the knowledge in the
SPECIALIST lexicon in UMLS and a database of synonyms. Various types of variants
are generated for a phrase word, including its acronyms, abbreviations, synonyms,
derivational variants, inflectional and spelling variants.
3. Candidate retrieval. Any string in Metathesaurus containing at least one of the variants
is retrieved as a candidate concept.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
55
4. Candidate evaluation. Each candidate is evaluated against the original phrase according to some weighting schema. The candidates are then ordered according to the mapping strength.
5. Mapping construction. Candidates involved in the phrase are combined to construct
complete mappings, and the strength of the mappings is computed using the same
schema in candidate evaluation. Mappings with the highest strength represent the best
interpretation of MetaMap of the original phrase.
Figure 4.2 shows an example of the output of MetaMap. The sentence in the figure is
processed by MetaMap and only part of the output is shown. The complete output is listed
in Appendix E. In this example, MetaMap identified a concept myocardial infarction in the
sentence, which is in correspondence to the candidate noun infarction. Therefore, the phrase
myocardial infarction is used as a candidate, while the original noun infarction is not included
in the new candidate set.
4.4
Representing candidates using features
We expect that candidates in the same semantic class will have similar behavior. Therefore,
the idea of the classification is to group together similar candidates. The similarity is characterized by syntactic relations, context information, and semantic types in UMLS. All features
are binary features, i.e., a feature takes value 1 if it is present; otherwise, it takes value 0.
Global syntactic relations Syntactic relations have been explored to group similar words
[Lin, 1998] and words of the same sense in word sense disambiguation [Kohomban and Lee,
2005].
Given the following sentences [Lin, 1998]:
A bottle of tezguino is on the table.
Everyone likes tezguino.
Tezguino makes you drunk.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
56
Sentence:
It found that the combined rate of myocardial infarction, stroke, or death was slightly lower in
the lower dose than in the higher dose groups at 3 months.
Output of MetaMap:
Phrase: “of myocardial infarction”
Meta Candidates (6)
1000 Myocardial Infarction [Disease or Syndrome]
861 Infarction [Finding,Pathologic Function]
861 Myocardial [Functional Concept]
805 MI <2> (Without) [Qualitative Concept]
789 MIS (Mullerian duct inhibiting substance) [Amino Acid, Peptide, or Protein,Hormone]
789 Myocardium [Tissue]
Meta Mapping (1000)
1000 Myocardial Infarction [Disease or Syndrome]
Phrase: “stroke”
Meta Candidates (1)
1000 Stroke (Cerebrovascular accident) [Disease or Syndrome]
Meta Mapping (1000)
1000 Stroke (Cerebrovascular accident) [Disease or Syndrome]
Phrase: “or”
Meta Candidates (0): <none>
Meta Mappings: <none>
Figure 4.2: Example of output of MetaMap
C HAPTER 4. C ORES OF SEMANTIC CLASSES
57
We make tezguino out of corn.
Lin tried to infer that tezguino is similar to beer, wine, etc., i.e., it is a kind of drink, by
comparing syntactic relations in which each word participates.
Kohomban and Lee [2005] determine the sense of a word in a context by observing a
subset of all syntactic relations in the corpus that the word participates in. The hypothesis is
that different instances of the same sense will have similar relations.
In our work, we need to group cores of the same semantic class. Such cores may participate
in similar syntactic relations while those of different classes will have different relations. For
example, intervention-cores often are subjects of sentences, while outcome-cores are often
objects.
Candidates in our task are phrases, instead of words as in [Lin, 1998] and [Kohomban
and Lee, 2005]. Thus, we extend their approaches of analyzing relations between two words
to extract relations between a word and a phrase. This is done by considering all relations
between a candidate noun phrase and other words in the sentence. To do that, we ignore
relations between any two words in the phrase when extracting syntactic relations. Any relation
between a word not in the phrase and a word in the phrase is extracted. We use the Minipar
parser [Lin, 1994] to get the syntactic relations between words. After a sentence is parsed,
we extract relevant syntactic relations from the output of the parser. A relation is represented
using a triple that contains two words (one of them is in the noun phrase and the other is not)
and the grammatical relation between them. Figure 4.3 shows relevant triples extracted from
a sentence. The output of Minipar on this sentence is shown in Appendix F . Because long
distance relations are considered, the relation between thrombolysis and increases is captured.
In the feature construction, a triple is taken as a feature. The set of all distinct triples is the
syntactic relation feature set in the classification.
Local context
Context of candidates is also important in distinguishing different classes. For
example, a disease-core may often have people with in its left context. However, it is very
unlikely that the phrase people with mortality will occur in the text.
As we mentioned, a sentence often contains several instances of semantic classes (hence
several cores) that we are interested in. Wide-window context is not of much use in differenti-
C HAPTER 4. C ORES OF SEMANTIC CLASSES
58
Sentence:
Thrombolysis reduces the risk of dependency, but increases the chance of death.
Candidates:
thrombolysis, dependency, death
Relations:
(thrombolysis subj–of increase), (thrombolysis subj–of reduce)
(dependency pcomp-n–of of)
(death pcomp-n–of of)
Figure 4.3: Example of dependency triples extracted from output of Minipar parser.
ating these cores. In our experiment, we considered the two words on both sides of a candidate
(stop words were excluded). When extracting context features, all punctuation marks were
removed except the sentence boundary. The window did not cross boundaries of sentences.
We evaluated two representations of context: with and without order. In the ordered case,
local context to the left of the phrase is marked by -LLL, that to the right is marked by RRR-.
Symbols -LLL and RRR- are used only to indicate the order of text. For the candidate dependency in Figure 4.3, the context features with order are: reduces-LLL, risk-LLL, RRR-increases,
and RRR-chance. The context features without order are: reduces, risk, increases, and chance.
This example shows a case where ordered context helps distinguish an intervention-core
from an outcome-core. If order is not considered, candidates thrombolysis and dependency
have overlapped context: reduces and risk. When taking order into account, they have no overlapped features at all – thrombolysis has features RRR-reduces and RRR-risk, while dependency
has features reduces-LLL and risk-LLL.
Domain features As described in the mapping to concepts step in the preprocessing, at the
same time of mapping text to concepts in UMLS, MetaMap also finds their semantic types.
Each candidate has a semantic type defined in the Semantic Network of UMLS. For example,
the semantic type of death is organism function, that of disability is pathologic function, and
that of dependency is physical disability. These semantic types are used as features in the
C HAPTER 4. C ORES OF SEMANTIC CLASSES
59
Table 4.1: Number of Instances of Cores in the Whole Data Set
Intervention-core
501
Disease-core Outcome-core
153
384
Total
1038
classification.
4.5
Data set
Two sections of CE were used in the experiments. A clinician labeled the text for interventioncores and disease-cores. Complete clinical outcomes are also identified. Using the annotation
as a basis, outcome-cores were labeled by the author. The number of instances of each class is
shown in Table 4.1.
Data analysis In our approach, the design of the features is intended to group similar cores
together. As a first step to verify how well the intention is captured by the features, we observe
the geometric structure of the data.
In the analysis, candidates are derived using the domain specificity measure p(c | n). Each
candidate is represented by a vector of dimensionality D, where each dimension corresponds
to a single feature. The feature set consists of syntactic features, ordered context, and semantic
types. We map the high-dimensional data space to a low-dimensional space using the locally
linear embedding (LLE) algorithm [Roweis and Saul, 2000] for easy observation. LLE maps
high-dimensional data into a single global coordinate system of low dimensionality by reconstructing each data point from its neighbors. The contribution of the neighbors, summarized
by the reconstruction weights, captures intrinsic geometric properties of the data. Because
such properties are independent of linear transformations that are needed to map the original high-dimensional coordinates of each neighborhood to the low-dimensional coordinates,
they are equally valid in the low-dimensional space. In Figure 4.4, the data is mapped to a
3-dimensional space (the coordinate axes in the figure do not have specific meanings as they
do not represent coordinates of real data). Candidates of the four classes (intervention-core,
C HAPTER 4. C ORES OF SEMANTIC CLASSES
60
disease-core, outcome-core, and other) are represented by (red) stars, (blue) circles, (green)
crosses, and (black) triangles, respectively. We can see that candidates in the same class are
close to each other, and clusters of data points are observed in the figure.
5
4
3
−5
2
1
0
0
5
−1
10
−2
−6
−4
−2
0
2
4
15
Figure 4.4: Manifold structure of data
4.6
The model of classification
Because our classification strategy is to group together similar cores and the cluster structure
of the data is observed, we chose a semi-supervised learning model developed by Zhu et al.
[2003] that explores the cluster structure of data in classification. The general hypothesis of
this approach is that similar data points will have similar labels.
A graph is constructed in this model. In the graph, nodes correspond to both labeled and unlabeled data points (candidates of cores), and an edge between two nodes is weighted according
to the similarity of the nodes. More formally, let (x1 , y1 ), . . . , (xl , yl ) be labeled data, where
C HAPTER 4. C ORES OF SEMANTIC CLASSES
61
YL = y1 , . . . , yl are corresponding class labels. Similarly, let (xl+1 , yl+1 ), . . . , (xl+u , yl+u ) be
unlabeled data, where YU = yl+1 , . . . , yl+u are labels to be predicted. A connected graph
G = (V, E) can be constructed, where the set of nodes V correspond to both labeled and unlabeled data points and E is the set of edges. The edge between two nodes i, j is weighted.
Weights wij are assigned to agree with the hypothesis; for example, using a radial basis function (RBF) kernel: wij = exp(−d2 (xi , xj )/σ 2 ), we can assign larger edge weights to closer
points in Euclidean space.
Zhu et al. developed two approaches of propagating labels from labeled data points to
unlabeled data points which have the same solution to the problem (the optimum solution is
unique). One of them follows closely the intuition of the propagation, while the other is defined
within a better framework. The first is described here to help understand the intuition of the
model, and the second is depicted because it is used in the experiment.
The iteration approach In the prediction, labels are pushed from labeled points through
edges to all unlabeled points using a probabilistic transition matrix, where larger edge weights
allow labels to travel through easier. The (l + u) × (l + u) probabilistic transition matrix T is
defined as [Zhu and Ghahramani, 2002]:
wij
Tij = Pl+u
k=1 wkj
where Tij is the probability moving from j to i. A label matrix B is a (l + u) × c matrix, where
c is the number of classes in the task, and each row represents the label probability distribution
of a data point.
In this problem setup, Zhu and Ghahramani proposed the label propagation algorithm:
1. Propagate B ← T B;
2. Row-normalize B to maintain the probability interpretation of the row;
3. Clamp the labeled data to keep the knowledge of originally labeled data;
4. Repeat from step 1 until B converges.
The label of a data point is determined by the largest probability in a row of B. It has been
proved that the algorithm converges. In fact, the solution can be directly obtained without
iterative propagation.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
62
Label propagation using Gaussian random fields In [Zhu et al., 2003], Zhu et al. formulated the intuitive label propagation approach as a problem of energy minimization in
the framework of Gaussian random fields, where the Gaussian field is over a continuous
state space, instead of over discrete label set. The idea is to compute a real-valued function
P
f : V → R on graph G that minimizes the energy function E(f ) = 21 i,j wij (f (i) − f (j))2 ,
where i and j correspond to data points in the problem. The function f = argmin f E(f )
determines the labels of unlabeled data points. This solution can be efficiently computed by
direct matrix calculation even for multi-label classification, in which solutions are generally
computationally expensive in other frameworks.
This approach propagates labels from labeled data points to unlabeled data points according to the similarity on the edges, thus it follows closely the cluster structure of the data in
prediction. We expect it to perform reasonably well on our data set. It is referred to as “SEMI”
in the following description.
4.7
Results and analysis
We use SemiL [Huang et al., 2006], an implementation of the algorithm using Gaussian random
fields in the experiment. SemiL provides different options for classification, among them some
are pertinent to our problem setting:
• Distance type. The distance between two nodes can be either Euclidean distance or
Cosine distance.
• Kernel type. The function used to assign weights on the edge. We use RBF in our
experiment. The Sigma value in the RBF kernel is set heuristically using labeled data
(Sigma is set to be the median of the distance from each data point in the positive class
to its nearest neighbour in the negative class [Jaakkola et al., 1999].). 2
• Normalization of the real-valued function f . It is designed to minimize the effect of
unbalanced data set in the classification. As our data is unbalanced, we turn on this
2
For heuristically set Sigma values in the thesis, several other Sigma values were used to verify the setting,
and the results show that the performance is stable.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
63
parameter to treat each class equally.
The performance of using Euclidean distance and Cosine distance in the similarity measure
is compared in the experiment in Section 4.7.4. Default values are used for the rest of the
parameters.
We first evaluate the performance of the semi-supervised model on different feature sets.
Then, we compare the two candidate sets obtained by using tf · idf and domain specificity
p(c | n), respectively. Finally, we compare the semi-supervised model to a supervised approach
to justify the usage of a semi-supervised approach in the problem.
In all experiments, the data set contains all candidates of cores. Unless otherwise mentioned, the result reported is achieved by using the candidate set derived by p(c | n), the feature
set of the combination of syntactic relations, ordered context, and semantic types, and the distance measure of cosine distance. The result of an experiment is the average of 20 runs. In
each run, labeled data is randomly selected from the candidate set, and the rest is unlabeled
data whose labels need to be predicted. We make sure all classes are present in labeled data. If
any class is absent, we redo the sampling. The evaluation of the semantic classes is very strict:
a candidate is given credit if it gets the same label as given by the annotator, and the tokens it
contains are exactly the same as marked by the annotator. Candidates that contain only some
of the tokens matching the labels given by the annotators are treated as the other class in the
evaluation.
4.7.1 Experiment 1: Evaluation of feature sets
This experiment evaluates different feature sets in the classification. As described in section
4.3, two options are used in the second step of preprocessing to pick up good candidates. Here,
as our focus is on the feature set we report only results on candidates selected by p(c | n). The
number of instances of each of the four target classes in the candidate set is shown in Table 4.2
(The performance of candidate selection will be discussed in subsection 4.7.2).
Figure 4.5 shows the accuracy of classification using different combinations of four feature
sets: syntactic relations, ordered context, un-ordered context, and semantic types. We set a
baseline by assigning labels to data points according to the prior knowledge of the distribution
C HAPTER 4. C ORES OF SEMANTIC CLASSES
64
Table 4.2: Number of Instances of Target Classes in the Candidate Set
Intervention-core
Disease-core Outcome-core
298
106
209
Others
Total
801
1414
0.8
0.7
Accuracy
0.6
0.5
0.4
baseline using class prior
baseline using majority label
rel
rel+orderco
rel+co
rel+orderco+tp
rel+co+tp
0.3
0.2
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Fraction of data used as labeled data
rel: syntactic relations
orderco: ordered context
co: no-order context
tp: semantic types
Figure 4.5: Classification Results of Candidates
of the four classes, which has accuracy of 0.395. Another choice of baseline is to assign the
label of the majority class, others in this case, to each data point, which produces an accuracy
of 0.567. However, all the three classes of interest have accuracy 0 according to this baseline.
Thus, this baseline is not very informative in this experiment.
It is clear in the figure that incorporating new kinds of features into the classification results
in a large improvement in accuracy. Only using syntactic relations (rel in the figure) as features,
C HAPTER 4. C ORES OF SEMANTIC CLASSES
65
the best accuracy is a little lower than 0.5, which is much higher than the baseline of 0.395. The
addition of ordered context (orderco) or no-order context features (co)) improved the accuracy
by about 0.1. Adding semantic type features (tp) further improved 0.1 in accuracy. Combining
all three kinds of features achieves the best performance. With only 5% of data as labeled
data, the whole feature set achieves an accuracy of 0.6, which is much higher than the baseline
of 0.395. Semantic type seems to be a very powerful feature set as it substantially improves
the performance on top of the combination of the other two kinds of features. Therefore, we
took a closer look at the semantic type feature set by conducting the classification using only
semantic types, and found that the result is even worse than using only syntactic relations. This
observation reveals interesting relations of the feature sets. In the space defined by only one
kind of features, data points may be close to each other, hence hard to distinguish. Adding
another kind sets apart data points in different classes toward a more separable position in the
new space. It shows that every kind of feature is informative to the task. The feature sets
characterize the candidates from different angles that are complementary in the task.
We also see that there is almost no difference between ordered and unordered context in
distinguishing the target classes, although ordered context seems to be slightly better when
semantic types are not considered.
4.7.2 Experiment 2: Evaluation of candidate sets
In the second step of preprocessing, one of two research options can be used to filter out some
bad nouns – using the tf · idf measure or the domain specificity measure p(c | n). This experiment compares the two measures in the core identification task. A third option using neither
of the two measures (i.e., skip the second step of preprocessing) is evaluated as the baseline.
The first three rows in Table 4.3 are numbers of instances remaining in the candidate set after
preprocessing. The last row shows the numbers of manually annotated true cores, which has
been listed in Table 4.1 and is repeated here for comparison. We analyze the classification
results using the candidate sets derived by tf · idf , domain specificity, and baseline to evaluate
the second step of preprocessing. Then, we compare the baseline to the manually annotated set
of cores to evaluate the first and third steps of preprocessing.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
66
Table 4.3: Number of Candidates in Different Candidate Sets
Measures
Intervention-core
Disease-core Outcome-core
Others
tf · idf
243
108
194
785
p(c | n)
298
106
209
801
baseline
303
108
236
1330
true cores
501
153
384
–
tf · idf , domain specificity vs. baseline As shown in Table 4.3, there are much fewer instances in the others class in the sets derived by tf ·idf and the probability measure as compared
to those derived by the baseline, which shows that the two measures effectively removed some
of the bad candidates of intervention-core, disease-core, and outcome-core. At the same time,
a small number of real cores were removed. Compared to the baseline method, the probability measure kept almost the same number of intervention-cores and disease-cores in the
candidate set, while omitting some outcome-cores. It indicates that outcome-cores are less
domain-specific than intervention-cores and disease-cores. Compared to the tf · idf measure,
more intervention-cores and outcome-cores were kept by the conditional probability measure,
showing that the probability measuring the domain-specificity of a noun better characterizes
the cores of the three semantic classes. The probability measure is also more robust than
the tf · idf measure, as tf · idf relies more on the content of the text from which it is calculated. For example, if an intervention is mentioned in many documents of the document set, its
tf · idf value can be very low although it is a good candidate of intervention-core.
The precision, recall, and F-score of the classification shown in Table 4.4 confirms the
above analysis. The domain specificity measure gets substantially higher F -scores than the
baseline for all the three classes that we are interested in, using different amounts of labeled
data. Compared to tf ·idf , the performance of the domain specificity measure is much better on
identifying intervention-core (note that p(c | n)picked up more real intervention-cores than tf ·
idf ), and slightly better on identifying outcome-cores, while the two are similar on identifying
disease-cores.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
67
Table 4.4: Results of Classification on Different Candidate Sets
INT: intervention-core DIS: disease-core OUT: outcome-core
labeled data
INT
DIS
OUT
1%
5%
10%
30%
60%
P
R
F
P
R
F
P
R
F
P
R
F
P
R
F
baseline
.44
.69
.53
.51
.83
.63
.53
.87
.66
.58
.90
.70
.59
.92
.72
tf · idf
.44
.62
.51
.52
.74
.61
.55
.77
.64
.59
.84
.69
.60
.87
.71
p(c | n)
.51
.65
.57
.60
.83
.69
.62
.86
.72
.65
.90
.75
.67
.91
.77
baseline
.16
.63
.25
.25
.68
.36
.31
.73
.43
.34
.84
.48
.35
.86
.49
tf · idf
.20
.55
.29
.31
.64
.41
.34
.70
.46
.39
.82
.53
.41
.86
.55
p(c | n)
.18
.56
.27
.30
.66
.41
.34
.73
.47
.39
.83
.53
.41
.87
.55
baseline
.22
.42
.28
.33
.53
.41
.39
.61
.48
.44
.66
.53
.46
.69
.55
tf · idf
.30
.43
.35
.43
.56
.49
.47
.61
.53
.53
.66
.59
.55
.70
.61
p(c | n)
.31
.46
.37
.43
.56
.49
.48
.62
.54
.54
.69
.60
.56
.71
.63
Baseline vs. the set of manually annotated cores As mentioned at the beginning of this
subsection, the baseline candidate set was derived by the first (PoS tagging) and third step
(mapping from nouns to concepts) in the preprocessing. As shown by Table 4.3, 62.3% of
manually annotated cores are kept in the baseline. We roughly checked about one-third of the
total true cores (manually annotated cores) in the data set and found that 80% of lost cores
are because MetaMap either extracted more or less tokens than marked by the annotator, or it
failed to find the concepts. 10% of missing cores are caused by errors of the PoS tagger, and
the rest are because some cores are not nouns.
4.7.3 Experiment 3: Comparison of the semi-supervised model and SVMs
In the semi-supervised model, labels propagate along high-density data trails, and settle down
at low-density gaps. If the data has this desired structure, unlabeled data can be used to help
learning. In contrast, a supervised approach only makes use of labeled data. This experiment
compares SEMI to a state-of-the-art supervised approach; the goal is to investigate how well
unlabeled data contributes to the classification using the semi-supervised model. We compare
C HAPTER 4. C ORES OF SEMANTIC CLASSES
68
the performance of SEMI to support-vector machines (SVMs) when different amounts of data
are used as labeled data.
Support vector machines In SVMs, the process of classification given a set of training examples is an optimization procedure searching for the optimal rule that predicts the label of
unseen examples with minimum errors. The goal of classification is to infer a rule from a sample of labeled training examples so that it recognize new examples with high accuracy. More
formally, the learner is given a training sample of n examples
(~x1 , y1 ), ..., (~xn , yn )
drawn according to an unknown but fixed distribution. Here ~x i are patterns, yi are labels or
targets. In the classification, a function needs to be found based on training data:
f : X → {+
− 1},
where the domain X is some nonempty set that the patterns ~xi are taken from, so that it will
correctly classify unseen examples.
The goal of SVMs is to find an optimal hyperplane so that examples on the same side of
the hyperplane will have the same label. SVMs learn decision functions:

 +1 : w
~ × ~x + b > 0
f (~x) = sgn((w
~ × ~x) + b) =
 −1 : otherwise.
(4.1)
Each function corresponds to a hyperplane in the feature space. The classification task is then
to determine on which side of the hyperplane a data point lies.
The optimal hyperplane that SVMs chose is the one with the largest margin. In the separable case, suppose we have a hyperplane that separates the positive examples from the negative
examples. Let d+ (d− ) be the shortest distance from the separating hyperplane to the closest
positive (negative) example. The margin of such a hyperplane is d + + d− . For the linear sepa~ 0 and a b0 , such that all positive training examples lie on one side of the
rable case, there is a w
hyperplane, while all negative examples lie on the other side. In general, there can be more
than one such hyperplanes, as shown in figure 4.6. Support vector machines choose the one
with the largest margin (H 0 in the figure).
C HAPTER 4. C ORES OF SEMANTIC CLASSES
69
H’
−
−
−
−
+
W
+
−
+
−
+
−b/w
Origin
+
−
H2
+
+
H1
Margin
Figure 4.6: Linear separating hyperplanes in two dimensions. The support vectors are marked
by squares.
In this experiment, we use OSU SVM [Ma et al., 2003], a toolbox for Matlab built on top
of LIBSVM [Chang and Lin, 2001]. LIBSVM is an implementation of SVMs. We use RBF
as the kernel method, and set the Sigma value heuristically using labeled data. SVM addresses
the problem of unbalanced data using a parameter, which assigns weights to each class in the
task. A class with larger weight will get more penalty when finding the optimum hyperplane.
We set the parameter according to the prior knowledge of the class distribution and give larger
weight to a class that contains less instances. Default values are used for other parameters.
Comparison of SEMI to SVMs As shown in Table 4.5, when there is only a small amount
of labeled data (less than 5% of the whole data set), which is often the case in real-world
applications, SEMI achieves much better performance than SVMs in identifying all the three
classes. For intervention-core and outcome-core, with 5% data as labeled data, SEMI outperforms SVMs with 10% data as labeled data. With less than 60% data as labeled data, the
performance of SEMI is either superior to or comparable to SVMs for intervention-core and
outcome-core. It shows that SEMI effectively exploits unlabeled data by following the manifold structure of the data. The promising results achieved by SEMI show the potential of
exploring unlabeled data in classification.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
70
Table 4.5: F-score of Classification Using Different Models
Candidate set: produced by p(c | n)(see Table 4.2)
INT: intervention-core DIS: disease-core OUT: outcome-core
labeled data
INT
DIS
OUT
1%
5%
10%
30%
60%
SEMI
.57
.69
.72
.75
.77
SVM
.33
.60
.68
.74
.77
SEMI
.27
.41
.47
.53
.55
SVM
.21
.38
.54
.62
.65
SEMI
.37
.49
.54
.60
.63
SVM
.07
.27
.44
.56
.62
Table 4.6: Accuracy using different distance measures.
Fraction of data as labeled data 10% 20%
30% 40% 50% 60%
Cosine distance
.647
.675
.687
.695
.701
.702
Euclidean distance
.341
.372
.405
.413
.410
.440
4.7.4 Experiment 4: Evaluation of distance measures
In the semi-supervised model, the cluster structure of the data is specified by the similarity of
data points. Therefore, the choice of distance measure affects the performance of the classification. In this experiment, we compare two distance measures: cosine distance and Euclidean
distance. Table 4.6 shows the classification accuracy using the two distance measures. The results show a large difference between them. Cosine distance is absolutely superior to Euclidean
distance in the classification task.
The σ value in the RBF kernel is a scale parameter of the distance between two data points.
Too-large a value can blur the distance between two well-separated points, while too-small a
value may improperly enlarge the gap between data points. If σ is within a reasonable range,
the performance of the classification will be relatively stable. Although parameter selection
C HAPTER 4. C ORES OF SEMANTIC CLASSES
71
was not a focus of the current work, we plot the results of using several different σ values in
the classification in Appendix C to have some sense of the its effect.
4.8
Related work
The task of named entity (NE) identification, similar to the core-detection task, involves identifying words or word sequence in several classes, such as proper names (locations, persons, and
organizations), monetary expressions, dates and times. NE identification has been an important
research topic ever since it was defined in MUC [MUC, 1995]. In 2003, it was taken as the
shared-task in CoNLL [Sang and Meulder, 2003]. Most statistical approaches use supervised
methods to address the problem [Florian et al., 2003; Chieu and Ng, 2003; Klein et al., 2003].
Unsupervised approaches have also been tried in this task. Cucerzan and Yarowsky [1999] use
a bootstrapping algorithm to learn contextual and morphological patterns iteratively. Collins
and Singer [1999] tested the performance of several unsupervised algorithms on the problem:
modified bootstrapping (DL-CoTrain) motivated by co-training [Blum and Mitchell, 1998], an
extended boosting algorithm (CoBoost), and the Expectation Maximization (EM) algorithm.
The results show that DL-CoTrain and CoBoost are superior to EM, while the two are almost
the same.
Much effort in entity extraction in the biomedical domain has gene names as the target.
Various supervised models including Naive Bayes, Support Vector Machines, Hidden Markov
Models have been applied [Ananiadou and Tsujii, 2003]. The work most related to our coreidentification in biomedical domain is that of Rosario and Hearst [2004], which extracts treatment and disease from MEDLINE and examines seven relation types between them using generative models and a neural network. They claim that these models may be useful when only
partially labeled data is available, although only supervised learning is conducted in the paper.
The best F-score of identifying treatment and disease obtained by using the supervised method
is .71. Another piece of work extracting similar semantic classes is in [Ray and Craven, 2001].
They report an F-score of about .32 for extracting proteins and locations, and an F-score of
about .50 for gene and disorder.
C HAPTER 4. C ORES OF SEMANTIC CLASSES
4.9
72
Summary
In this chapter, we identified an important property of semantic classes – the core and explained
its role in matching a question to its answer. Then, we proposed a novel approach to automatically identify and classify cores of instances of semantic classes in scenario descriptions. A
semi-supervised learning method was explored to reduce the need for manually annotated data.
In this approach, candidates of cores were first extracted from the text. We took two options
to obtain a better candidate set by removing noise from the original set: tf · idf was used to find
informative nouns, while a probability measure was to find domain-specific nouns. The results
show that both measures effectively remove some noise, while the probability measure better
captures the characteristics of cores. To do the classification, we designed several types of
features and represented each candidate with the syntactic relations in which it participates, its
context, and its semantic type, with the goal that candidates with similar representations are in
the same class. Our experimental results show that syntactic relations work well together with
other types of features. In the classification, a semi-supervised model that explores the manifold structure of the data was applied. The results show that the features characterize the cluster
structure of the data, and unlabeled data is effectively used. We compared the semi-supervised
approach to a state-of-the-art supervised approach, and showed that the performance of the
semi-supervised approach is much better when there is only a small amount of labeled data,
and performance of the two are comparable even when 60% of data are used as labeled data.
Our approach does not require prior knowledge of semantic classes, and it effectively exploits unlabeled data. The promising results achieved show the potential of semi-supervised
models that explore the cluster structure of data in similar tasks. Features of syntactic relations
and local context are general and can be used directly in tasks in other domain. The semantic type features make use of knowledge in UMLS, which is specific to medical domain. For
tasks that have domain-specific knowledge bases like UMLS, similar features can be generated
easily. For a domain without such knowledge base, the hierarchical information in WordNet
can be used as a replacement, although it would be more difficult as the level of generalization
needs to be determined.
A difficulty of using this approach, however, is in detecting boundaries of the targets. A
C HAPTER 4. C ORES OF SEMANTIC CLASSES
73
segmentation step that pre-processes the text is needed. This will be our future work, in which
we aim to investigate approaches that perform the segmentation precisely.
As a final point, we want to emphasize the difference between cores and named entities.
While the identification of NEs in a text is an important component of many tasks including
question answering and information extraction, its benefits are constrained by its coverage.
Typically, it is limited to a relatively small set of classes, such as person, time, and location.
However, in sophisticated applications, such as the non-factoid medical question answering
that we consider, NEs are only a small fraction of the important semantic units discussed in
documents or asked about by users. As shown by the examples in this chapter, cores of clinical
outcomes are often not NEs. In fact, many semantic roles in scenarios and events that occur in
questions and documents do not contain NEs at all. For example, the test method in diagnosis
scenarios, the means in a shipping event, and the manner in a criticize scenario may all have
non-NE cores. Therefore, it is imperative to identify other kinds of semantic units besides NEs.
Cores of semantic classes is one such extension that consist of a more diverse set of semantic
units that goes beyond simple NEs.
Chapter 5
Polarity of Clinical Outcomes
One of the major concerns in patient treatment is the clinical outcomes of interventions in
treating diseases: are they positive, negative or neutral? This polarity information is an inherent
property of clinical outcomes. An example of each type of polarity taken from CE is shown
below.
Positive: Thrombolysis reduced the risk of death or dependency at the end of the studies.
Negative: In the systematic review, thrombolysis increased fatal intracranial haemorrhage compared
with placebo.
Neutral: The first RCT found that diclofenac plus misoprostol versus placebo for 25 weeks produced
no significant difference in cognitive function or global status.
Sentences that do not have information on clinical outcomes form another group: no outcome.
No outcome: We found no RCTs comparing combined pharmacotherapy and psychotherapy with
either treatment alone.
Polarity information is crucial to answer questions related to clinical outcomes. We have to
know the polarity to answer questions about benefits and harms of an intervention. In addition,
knowing whether a sentence contains a clinical outcome can help filter out irrelevant information in answer construction. Furthermore, information on negative outcomes can be crucial in
clinical decision making.
74
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
75
In this chapter, we discuss the problem of automatically identifying outcome polarity in
medical text [Niu et al., 2005]. More specifically, we focus on detecting the presence of a
clinical outcome in medical text, and, when an outcome is found, determining whether it is
positive, negative, or neutral1 . We observe that a single sentence in medical text usually describes a complete clinical outcome. As a result, we perform sentence-level analysis in our
work.
5.1
Related work
The problem of polarity analysis is also considered as a task of sentiment classification [Pang
et al., 2002; Pang and Lee, 2004] or semantic orientation analysis [Turney, 2002]: determining
whether an evaluative text, such as a movie review, expresses a “favorable” or “unfavorable”
opinion. All these tasks are to obtain the orientation of the observed text on a discussion topic.
They fall into three categories: detection of the polarity of words, sentences, and documents.
Among them, as Yu and Hatzivassiloglou [2003] pointed out, the problem at the sentence level
is the hardest one.
Turney [2002] has employed an unsupervised learning method to provide suggestions on
documents as thumbs up or thumbs down. The polarity detection is done by averaging the semantic orientation (SO) of extracted phrases (phrases containing adjectives or adverbs) from
a text. The document is tagged as thumbs up if the average of SO is positive, and otherwise
is tagged as thumbs down. The SO is calculated by the difference in mutual information between an observed phrase and the positive word excellent and mutual information between the
observed phrase and the negative word poor. Documents are classified as either positive or
negative; no neutral position is allowed.
In more recent work, Whitelaw et al. [2005] explore appraisal groups to classify positive
and negative documents. Similar to phrases used in Turney’s work, appraisal groups consist of
coherent words that together express the polarity of opinions, such as “extremely boring”, or
“not really very good”. Instead of calculating the mutual information, a lexicon of adjectival
1
This part of the work was carried out in collaboration with Xiaodan Zhu and Jane Li. They participated in
the manual annotation. Xiaodan Zhu collected the BIGRAMS features, Jane Li collected the SEMANTIC TYPES.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
76
appraisal groups (groups headed by an appraising adjective) is constructed semi-automatically.
These groups are used as features in a supervised approach using SVMs to detect the sentiment
of a document.
Pang et al. [2002] also deal with the task at document level. The sentiment classification
problem were treated as a text classification issue and a variety of machine learning techniques were explored to classify movie reviews into positive and negative. Three classification
strategies, Naive Bayes, maximum entropy classification, and support vector machines, were
investigated, and a series of lexical features were employed on these classification strategies in
order to find effective features. Pang et al. found that machine learning techniques can always
outperform a human-generated baseline; among the three classification strategies, support vector machines perform the best and the Naive Bayes tends to be the worst; unigrams are the
most effective lexical feature and indispensable compared with the other alternatives.
The main part of Yu and Hatzivassiloglou’s work [Yu and Hatzivassiloglou, 2003] is at
the sentence level, and is hence most closely related to our work. They first separate facts
from opinions using a Bayesian classifier. Various features derived from observing semantic
orientation of words are tried in this step. After opinion sentences are identified, they then use
an unsupervised method to classify opinions into positive, negative, and neutral by evaluating
the strength of the orientation of words contained in a sentence. A gold standard is built for
evaluation, which includes 400 sentences labeled by one judge. On the task of distinguishing
opinions from facts, the best performance is recall=0.92, precision=0.70 for the opinion class.
The performance is much worse for the fact class. The best recall and precision obtained are
0.13 and 0.42. The unsupervised approach of detecting polarity of sentences achieves 0.62
accuracy.
The polarity information we are observing relates to clinical outcomes instead of the personal opinions studied by the work mentioned above. Therefore, we expect differences in the
expressions and the structures of sentences in these two areas. For the task in the medical domain, it will be interesting to see if domain knowledge will help. These differences lead to new
features in our approach.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
5.2
77
A supervised approach for clinical outcome detection and
polarity classification
As discussed in Section 5.1, various supervised models have been used in sentiment classification. At document-level, SVMs perform better than other models and achieve promising
results [Pang et al., 2002]. In sentence-level analysis, Yu and Hatzivassiloglou [2003] use a
Bayesian classifier to distinguish facts from opinions. The results for the fact class are not
very satisfactory, which indicates that the task at sentence level may be more difficult. Since
SVMs have been shown also very effective in many other classification tasks, in our work, we
investigate SVMs in sentence-level analysis to detect the presence of a clinical outcome and
determine its polarity.
In our approach, each sentence as a data point to be classified is represented by a vector of
features. In the feature set, we use words themselves as they are very informative in related
tasks such as sentiment classification and topic categorization. In addition, we use contextual
information to capture changes described in clinical outcomes, and use generalized features
that represent groups of concepts to build more regular patterns for classification.
We use binary features in most of the experiments except for the frequency feature in one
of our experiments. When a feature is present in a sentence, it has a value of 1; otherwise, it
has a value of 0. Among the features in our feature set, UNIGRAMS and BIGRAMS have been
used in previous sentiment classification tasks, and the rest are new features that we developed.
5.2.1 Unigrams
A sentence is composed of words. Distinct words (unigrams) can be used as the features of
a sentence. In previous work on sentiment classification [Pang et al., 2002; Yu and Hatzivassiloglou, 2003], unigrams are very effective. Following this work, we also take unigrams as
features. We use unigrams occurring more than 3 times in the data set in the feature set, and
they are called UNIGRAMS in the following description.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
78
5.2.2 Context features
Our observation is that outcomes often express a change in a clinical value [Niu and Hirst,
2004]. In the following example, mortality was reduced.
(23) In these three postinfarction trials ACE inhibitor versus placebo significantly reduced
mortality, readmission for heart failure, and reinfarction.
The polarity of an outcome is often determined by how a change happens: if a bad thing
(e.g., mortality) was reduced, then it is a positive outcome; if a bad thing was increased,
then the outcome is negative; if there is no change, then we get a neutral outcome. We tried
to capture this observation by adding context features – BIGRAMS, two types of CHANGE
PHRASES ( MORE / LESS
B IGRAMS
features and POLARITY- CHANGE features), and NEGATIONS.
Bigrams (two adjacent words) are also used in sentiment classification. In that
task, they are not so effective as UNIGRAMS. When combined with UNIGRAMS, they do not
improve the classification accuracy [Pang et al., 2002; Yu and Hatzivassiloglou, 2003]. However, in our task, the context of a word in a sentence that describes the change in a clinical value
is important in determining the polarity of a clinical outcome. Bigrams express the patterns of
pairs, and we expect that they will capture some of the changes. Therefore, they are used in
our feature set. As with UNIGRAMS, bigrams with frequency greater than 3 are extracted and
referred to by BIGRAMS.
C HANGE PHRASES
We developed two types of new features to capture the trend of changes
in clinical values. The collective name CHANGE PHRASES is used to refer to these features.
To construct these features, we manually collected four groups of words by observing several sections in CE: those indicating more (enhanced, higher, exceed, ...), those indicating less
(reduce, decline, fall, ...), those indicating good (benefit, improvement, advantage, ...), and
those indicating bad (suffer, adverse, hazards, ...).
• MORE / LESS features. This type of feature emphasizes the effect of words expressing
“changes”. The way the features are generated is similar to the way that Pang et al.
[2002] add negation features. We attached the tag MORE to all words between the
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
79
more-words and the following punctuation mark, or between the more-words and
another more(less) word, depending on which one comes first. The tag LESS was
added similarly. This way, the effect of the “change” words is propagated.
(24) The first systematic review found that β blockers significantly reduced LESS
the LESS risk LESS of LESS death LESS and LESS hospital LESS admissions
LESS.
(25) Another large rct (random clinical trial) found milrinone versus placebo increased
MORE mortality MORE over MORE 6 MORE months MORE.
• POLARITY- CHANGE features. This type of feature addresses the co-occurrence of
more/less words and good/bad words, i.e., it detects whether a sentence expresses
the idea of “change of polarity”. We used four features for this purpose: MORE
GOOD , MORE BAD , LESS GOOD ,
and LESS BAD. As this type of features aims for
the “changes” instead of “propagating the change effect”, we used a smaller window
size to build these features. To extract the first feature, a window of four words on
each side of a more-word in a sentence was observed. If a good-word occurs in this
window, then the feature MORE GOOD was activated (its value is set to 1). The other
three features were activated in a similar way.
N EGATIONS
Most frequently, negation expressions contain the word no or not. We observed
several sections of CE and found that not often does not affect the polarity of a sentence, as
shown in the following examples, so it is not included in the feature set.
(26) However, disagreement for uncommon but serious adverse safety outcomes has not
been examined.
(27) The first RCT found fewer episodes of infection while taking antibiotics than while
not taking antibiotics.
(28) The rates of adverse effects seemed higher with rivastigmine than with other anticholinesterase drugs, but direct comparisons have not been performed.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
80
The case for no is different: it often suggests a neutral polarity or no clinical outcome at all:
(29) There are no short or long term clinical benefits from the administration of nebulised
corticosteroids . . .
(30) One systematic review in people with Alzheimer’s disease found no significant benefit
with lecithin versus placebo.
(31) We found no systematic review or RCTs of rivastigmine in people with vascular dementia.
We develop the NEGATION features to take into account the evidence of the word no. To extract
the features, all the sentences in the data set are first parsed by the Apple Pie parser [Sekine,
1997] to get phrase information. Then, in a sentence containing the word no, the noun phrase
containing no is extracted. Every word in this noun phrase except no itself is attached by a NO
tag.
5.2.3 Semantic types
Using category information to represent groups of medical concepts may relieve the data
sparseness problem in the learning process. For example, we found that diseases are often
mentioned in clinical outcomes as bad things:
(32) A combined end point of death or disabling stroke was significantly lower in the
accelerated-t-PA group . . .
Thus, all names of specific diseases in the text are replaced with the tag DISEASE.
Intuitively, the occurrences of semantic types, such as pathologic function and organism
function, may be different in different polarity of outcomes, especially in the no outcome class
as compared to the other three classes. To verify this intuition, we collect all the semantic types
in the data set and use each of them as a feature. They are referred to as SEMANTIC TYPES.
Thus, in addition to the words contained in a sentence, all the medical categories mentioned in
a sentence are also considered.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
81
The Unified Medical Language System (UMLS) is used as the domain knowledge base for
extracting semantic types of concepts. The software MetaMap [Aronson, 2001] is incorporated
for mapping concepts to their corresponding semantic types in the UMLS Metathesaurus.
5.3
Experiments
We carried out several experiments on two text sources: CE and Medline abstracts. Compared to CE text, Medline has a more diverse writing style as different abstracts have different
authors. The performance of the supervised classification approach on the two sources is compared to find out if there is any difference. We believe that these experiments will lead to better
understanding of the polarity detection task.
5.3.1 Outcome detection and polarity classification in CE text
Using CE as the text source, we evaluate a two-way classification task of distinguishing positive from negative outcomes, and the four-way classification of positive, negative, neutral
outcomes, and no outcomes.
Positive vs. negative polarity
Experimental setup In this experiment, we have two target classes: positive outcomes and
negative outcomes. The training and test sets were built by collecting sentences from different
sections in CE; 772 sentences were used, 500 for training (300 positive, 200 negative), and 272
for testing (95 positive, 177 negative). All examples were labeled manually by the author.
We used the SVMlight implementation of SVMs [Joachims, 2002] to perform the classification and used the default values for the parameters.
Results and analysis Features used in the experiment are listed in the left-most column in
Table 5.1. We construct features in two ways: using presence of a feature, a binary feature
indicates whether a feature is present or not; and using frequency of a feature, the count of the
number of occurrences of a feature in the sentence. The accuracies achieved by presence of
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
82
Table 5.1: Accuracy of positive/negative classification using a linear kernel in CE
Features
Presence (%)
baseline
65.1
UNIGRAMS
89.0
UNIGRAMS
with DISEASE
90.1
UNIGRAMS
with MORE / LESS
91.5
UNIGRAMS
with DISEASE and MORE / LESS
92.7
features using a linear kernel (the default choice of kernels) are listed in Table 5.1. Frequency
of features produces approximately the same results.
The baseline is to assign the negative label to all test samples as it is more frequent in the
test set, which has the accuracy of 65.1%. As shown in the table, combining features achieves
an accuracy as high as 92.7%. Using a more general category DISEASE instead of specific
diseases has a positive effect on the classification. It is clear in the table that the MORE / LESS
features improve the performance. Compared to using only UNIGRAMS, the combined feature
set improves the accuracy by 0.037. The DISEASE and MORE / LESS features both contribute to
distinguishing positive from negative classes.
A non-linear kernel RBF (exp(−d2 (xi , xj )/σ 2 )) was also tested in SVMs. Using the feature
set of presence of combining UNIGRAMS with DISEASE and MORE / LESS, the accuracy of the
classification obtained with several σ values is shown in Appendix H. When σ is large, the
performance is not very sensitive to its change, and becomes relatively stable.
Four-way classification
Experimental setup The data set of sentences in all the four classes was built by collecting
sentences from different sections in CE (sentences were selected so that the data set is relatively
balanced). The number of instances in each class is shown in Table 5.2. The data set is labeled
manually by three graduate students, and each sentence is labeled by one of them. We used
the OSU SVM package [Ma et al., 2003] with an RBF kernel for this experiment. The σ value
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
83
Table 5.2: Number of instances in each class (CE)
Positive
Negative
Neutral
No-outcome
Total
472
338
250
449
1509
was set heuristically using training data. Default values were used for other parameters in the
package.
0.8
0.78
0.76
0.74
Accuracy
0.72
0.7
0.68
0.66
Unigrams
Unigrams+Bigrams
Unigrams+Bigrams+ChangePhrases
Unigrams+Bigrams+ChangePhrases+Negations
Unigrams+Bigrams+ChangePhrases+Negations+SemanticTypes
0.64
0.62
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Fraction of Training Data Used
Figure 5.1: Accuracy of classification using different fractions of training data
Results and analysis
We first randomly select 20% of the whole data set as the test set (301 sentences), and use the
rest (1208 sentences) as the training set. In the training process, we gradually add training samples until all of them are included, and observe the performance on the test set. The results are
shown in Figure 5.1. As the figure indicates, accuracy goes up as more training data is used, and
when more features are added. The complete feature set performs consistently the best. These
results match our intuition that context information (BIGRAMS and
CHANGE PHRASES )
and
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
84
Table 5.3: Results of the four-way classification with different feature sets in CE
Features
Accuracy (%)
Relative Error Reduction (%)
(to Unigrams)
(1)U NIGRAMS
76.9
–
(1)+(2)B IGRAMS
79.4
10.8
(1)+(2)+(3)C HANGE PHRASES
79.6
11.7
(1)+(2)+(3)+(4)N EGATIONS
79.6
11.7
(1)+(2)+(3)+(4)+(5)SEMANTIC TYPES
80.6
16.0
generalizations (SEMANTIC TYPES) are important factors in detecting the polarity of clinical
outcomes.
Table 5.3 shows the results of the five feature sets used for classification. The accuracy is
the average of 50 runs of the experiment. In each run, 20% of the data is selected randomly
as the test set, and the rest is used as the training set. With just UNIGRAMS as features, we
get 76.9% accuracy, which is taken as the baseline. The addition of BIGRAMS in the feature
set results in an increase of about 2.5% in accuracy, which corresponds to 10.8% of relative
error reduction. CHANGE PHRASES lead to a very small improvements and NEGATIONS do
not improve the performance on top of BIGRAMS. This result seems to be different from the
previous experiment of positive/negative classification, where the MORE / LESS features reduce
the error rate. Note that CHANGE PHRASES intend to capture the impact of context, and bigrams also contain context information. It could be that some effect of CHANGE PHRASES has
already been captured by bigrams. Also, since the target classes are different in the two tasks,
CHANGE PHRASES
may be more important in distinguishing positive from negative outcomes.
The SEMANTIC TYPES features further improve the performance on top of the combination of
other features, which shows that generalization is helpful.
Which class is the most difficult to detect, and why? To answer these questions, we further
examine the errors in every class. The precision, recall and F-score of each class are shown in
Table 5.4 (it is the result of one run of the experiment). It is clear in the table that the negative
class has the lowest precision and recall. A lot of errors occur in distinguishing negative from
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
85
Table 5.4: Classification results of each class on CE data
Positive
Negative
Neutral
No Outcome
Precision (%)
86.8
73.1
79.2
76.8
Recall (%)
83.2
73.1
76.0
82.0
F-score (%)
85.0
73.1
77.6
79.3
no-outcome classes. We studied the incorrectly classified sentences and found some interesting
cases. Some of the errors are because descriptions of diseases in the no-outcome class are often
identified as negative. These sentences are difficult in that they contain negative expressions
(e.g., increased risk), yet do not belong to the negative class:
(33) Lewy body dementia is an insidious impairment of executive functions with Parkinsonism, visual hallucinations, and fluctuating cognitive abilities and increased risk of
falls or autonomic failure.
Negative samples are sometimes assigned a positive label when a sentence has phrasings
that seem to contrast, as shown in the following example:
(34) The mean increase in height in the budesonide group was 1.1 cm less than in the
placebo group (22.7 vs 23.8 cm, P= 0005); . . .
In this sentence, the clinical outcome of impaired growth is expressed by comparing height
increase in two groups, which is less explicit and hard to capture.
5.3.2 Outcome detection and polarity classification in Medline
With Medline abstracts, we evaluate two tasks: the first one is two-way classification that
aims to detect the presence of clinical outcomes. In this task, a sentence is classified into two
classes: containing a clinical outcome or not. The second task is the four-way classification,
i.e., identifying whether an outcome is positive, negative, neutral, or a sentence does not contain
an outcome.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
86
Table 5.5: Number of instances in each class (Medline)
Positive
Negative
Neutral
No-outcome
Total
469
122
194
1513
2298
Table 5.6: Results of two-way and four-way classification with different feature sets (Medline)
RER=Relative Error Reduction (compared to unigrams)
Features
two-way
four-way
Accuracy (%)
RER (%)
Accuracy (%)
RER (%)
(1) U NIGRAMS
80.1
–
75.5
–
(1)+(2) B IGRAMS
81.7
8.0
77.4
7.8
82.0
9.5
77.6
8.6
81.9
9.0
77.6
8.6
82.5
12.1
78.3
11.4
(1)+(2)+(3) C HANGE
PHRASES
(1)+(2)+(3)+(4) N EGATIONS
(1)+(2)+(3)+(4)+(5) SEMANTIC
TYPES
Experimental setup We collected 197 abstracts from Medline that were cited in CE. The
number of sentences in each class is listed in Table 5.5. The data set was annotated with the
four classes of polarity information by two graduate students. Each single sentence is annotated
by one of them. In this experiment, again, 20% of the data was randomly selected as test set
and the rest was used as the training data. The averaged accuracy was obtained from 50 runs.
We used the same SVM package as in Section 5.3.1 for this experiment, parameters were set
in the same manner.
Results and analysis
Results of the two tasks are shown in Table 5.6. Not surprisingly, the
performance on the two-way classification is better than on the four-way task. For both tasks,
we see a similar trend in accuracy as in CE text (see Table 5.3). The accuracy goes up as
more features are added, and the complete feature set has the best performance. Compared
to UNIGRAMS, the combination of all features significantly improves the performance in both
tasks (paired t-test, p values < 0.0001). With just UNIGRAMS as features, we get 80.1% accu-
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
87
racy for the two-way task. The addition of BIGRAMS in the feature set results in a decrease of
1.6% in the error rate, which corresponds to 8.0% of relative error reduction as compared to
UNIGRAMS .
Similar improvements are observed in the four-way task. The SEMANTIC TYPES
features also slightly reduce the error rate.
Compared to the results on CE text in Table 5.3, the four-way classification task tends to be
more difficult on Medline text. This can be observed by comparing the improvement of adding
all other features to UNIGRAMS. As we mentioned in section 5.3, Medline abstracts have a
more diverse writing style because they are written by different authors. This could be a factor
that makes the classification task more difficult. However, the general performance of features
on Medline abstracts and CE text is similar, which shows that the feature set is relatively robust.
In our outcome detection and polarity classification task, UNIGRAMS are very effective
features, as has been previously shown in the context of sentiment classification problems.
This shows that information in words is very important for the polarity detection task. Context
information represented by BIGRAMS and CHANGE PHRASES is also valuable in our task (see
Table 5.1, Table 5.3, and Table 5.6). The effectiveness of BIGRAMS is different from the
results obtained by Pang et al. [2002] and Yu and Hatzivassiloglou [2003]. In their work,
adding bigrams does not make any difference in the accuracy, or even is slightly harmful in
some cases. This indicates the difference in the expression of polarity in clinical outcomes and
the polarity in opinions. Generalization features (DISEASE in Table 5.1, SEMANTIC TYPES in
Table 5.3 and Table 5.6) are also helpful in our task.
5.4
Discussion
The performance bottleneck in polarity classification As described in Section 5.1, supervised approaches have been used in sentiment classification. Features used in these approaches
usually include: n-grams, PoS tags, and features based on words with semantic orientations
(e.g., adjectives such as good, bad). In all such studies, a common observation is that unigrams are very effective, while adding more features does not gain much.
• In the task of detecting polarity of documents [Pang et al., 2002], the best performance
is obtained using unigrams.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
88
• In the sentence-level opinion/fact classification task [Yu and Hatzivassiloglou, 2003],
as described in section 5.1, various features based on semantic orientation of words
are tried, including counts of semantically oriented words, the polarity of the head
verbs and the average semantic orientation score of the words in the sentence. A gold
standard set is built which includes 400 sentences labeled by one judge. In the opinion
class, the only result better than the performance of unigrams is obtained by combining
all features, which results in only 0.01 improvement in precision. Similarly, not much
is achieved by adding all other features in detecting facts.
• In [Whitelaw et al., 2005], the best performance of the approach is achieved by the
combination of unigrams with the appraisal groups, which is 3% higher in accuracy
than using unigrams alone.
From all this work, we observe a performance bottleneck problem in the polarity classification task: various features have been developed; however, adding more features does not gain
much in classification accuracy, and it may even hurt the performance. In our task, although
the context and generalization features significantly improve the performance compared to unigrams, we observe a similar performance bottleneck problem.
Analysis of the problem The bottleneck problem shows that additional features have much
overlap with unigram features, and they may add noise to the classification.
We further analyzed the data, and found that most words in a sentence do not contribute
to the classification task. Instead, they can be noise that cannot be removed by adding more
features. This could be a crucial reason of the bottleneck discussed above.
To verify this hypothesis, we conducted some experiments on the Medline data set of 2298
sentences used in Section 5.3.2. From each sentence in the data set, we manually extract
some words that fully determine the polarity of the sentence. We refer to these words by
extractions in the following description. For those sentences that do not contain outcomes,
nothing is extracted. The following examples are some sentences with different polarity and
the extractions from them. These extractions form another data set, which we call the extraction
set.
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
89
Sentence:
Treatment with reperfusion therapies and achievement of TIMI 3 flow are associated with increased
short- and medium-term survival after infarction.
Extraction:
increased short- and medium-term survival
Sentence:
In all three studies, a significant decrease in linear growth occurred in children treated with beclomethasone compared to those receiving placebo or non-steroidal asthma therapy.
Extraction:
decrease in linear growth occurred
Sentence:
The doxazosin arm, compared with the chlorthalidone arm, had a higher risk of stroke.
Extraction:
a higher risk of stroke
Sentence:
Prednisolone treatment had no effect on any of the outcome measures.
Extraction:
no effect
Sentence:
There was no significant mortality difference during days 0-35, either among all randomised patients
or among the pre-specified subset presenting within 0-6 h of pain onset and with ST elevation on
the electrocardiogram in whom fibrinolytic treatment may have most to offer.
Extraction:
no significant mortality difference
We performed the four-way classification task on this extraction set. We constructed UNI GRAMS
feature based on the extraction set and used them in the classification. Using 80% of
the data as the training data and the rest as the test data, we achieved an accuracy of 93.3%,
which is much higher than the accuracy of the four-way classification task on the original
sentence set (75.5%).
The fact that we do not extract any words from no-outcome sentences may make the task
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
90
easier. Therefore, we removed from the extraction set all sentences that do not contain an
outcome, and reran the experiment. This task has three target classes: positive, negative or
neutral. We obtained an accuracy of 82.2%. However, performing the three-way classification
on the original sentence set only achieves 70.7% accuracy.
The results clearly show that irrelevant words actually introduce a lot of noise in the polarity
detection task. Therefore, a new direction of research on the task is to conduct feature selection
to remove words that do not contribute to the classification.
A possible solution We took a closer look at the extraction set and found that the extractions usually form a sequence or several sequences in a sentence. Because Hidden Markov
Model and Conditional Random Fields are effective models for sequence detection, they will
be explored in the future work of this research.
5.5
Summary
In this chapter, we discussed an approach of identifying an inherent property of clinical outcomes – their polarity. Polarity information is important to answer questions related to clinical
outcomes. We explored a supervised approach to detect the presence of clinical outcomes and
their polarity. We analyzed this problem from various aspects:
• We developed features to represent context information and explored domain knowledge to get generalized features. The results show that adding these features significantly improves the classification accuracy.
• We showed that the feature set has consistent performance on two different text sources,
CE and Medline abstracts.
• We evaluated the performance of the feature set on different subtasks of the outcome
detection to understand how difficult each subtask is.
• We compared outcome polarity detection to sentiment classification according to different performance of context features on the two tasks. We found that bigram features
C HAPTER 5. P OLARITY OF C LINICAL O UTCOMES
91
have almost no effect on the sentiment classification task, while they improve the classification accuracy of identifying presence and polarity of clinical outcomes.
• We identified a performance bottleneck problem in the polarity classification task using
a supervised approach. In both the sentiment classification and the outcome polarity
detection, we observed that adding more features on top of the unigram features does
not lead to major improvement in accuracy. We found a crucial reason for this – the
noise in the feature set is not removed by adding more features.
• We proposed to use Hidden Markov Model or Conditional Random Fields to conduct
feature selection and thus to remove noise from the feature set.
Chapter 6
Sentence Extraction using Outcome
Polarity
As we have addressed in Section 1.5, a crucial characteristic of NFQA is to identify multiple
pieces of relevant information to construct answers. In the two previous chapters, we discussed
properties of semantic classes that are important for detecting the relevance of a piece of information. In this chapter, we investigate the problem of relevance detection using one of the
properties: information on the polarity of clinical outcomes, which is discussed in Chapter 5
[Niu et al., 2006].
6.1
Related work
The work most similar to ours is the multi-perspective question answering (MPQA) task, in
which Stoyanov et al. [2005] argue that presence of opinions should be identified to find the
correct answer for a given question. Some preliminary results are presented to support this
claim. Stoyanov et al. [2005] manually created a corpus of opinion and fact questions and
answers, OpQA, which consists of 98 documents that appeared in the world press. The documents cover four general topics: President Bush’s alternative to the Kyoto protocol; the US
annual human rights report; the 2002 coup d’etat in Venezuela; and the 2002 elections in Zimbabwe and Mugabe’s reelection. Each topic is covered by between 19 and 33 documents. For
each topic, there are 3 to 4 opinion questions, and there are 15 questions in total for all topics.
92
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
93
In their answer rank experiments, each sentence in the whole document set is taken as a potential answer to a question. Sentences are first ranked by an information retrieval algorithm
based on tf · idf of words in the sentences (step1). Then all fact sentences (sentences that do
not express opinions) are removed by some subjectivity filters that distinguish between facts
and opinions (step2). In the evaluation, the rank of the first answer to each question in the
ranked list after step1 is compared to the rank of the first answer in the list after step2. The
mean reciprocal rank (MRR) is used as the evaluation metric. Their results show that the MRR
value after step2 is higher than the value after step1.
The results indicate the value of using subjectivity filters in MPQA. The experiment also inspires more thoughts on similar problems. For example, this experiment takes a single sentence
as a potential answer, which does not meet very well the needs of drawing on multiple pieces of
information in constructing answers to opinion-related questions. Moreover, the strategy of filtering out irrelevant information by removing any sentence that does not express opinions could
be too simple for the complex QA task. In our work, we address these problems by exploiting multi-document summarization techniques to find sentences that are relevant/important for
answering questions about clinical outcomes, such as “What are the effects of intervention A
on disease B?”. More specifically, the problem is: after a set of relevant documents has been
retrieved, how can we locate constituents of the answer in these documents? 1
We believe summarization techniques are suitable for our task for two main reasons. First,
simply filtering out any information that does not contain an outcome is not appropriate in
answer construction. As we discussed in Section 1.1, different outcomes may be present in
different patient groups or clinical trials. Therefore, besides information on clinical outcomes,
explanation on conditions of the patient groups or the clinical trials can be very important as
well. Moreover, not every piece of clinical outcome is important; unimportant outcomes should
be discarded. Second, the goal of the summarization task is to find important information
with the smallest redundancy, which agrees with that of answer construction in non-factoid
QA. The connection between QA and summarization is attracting more attention in the text
summarization community. In 2003, the document understanding conferences (DUC) started
1
This part of the work was carried out in collaboration with Xiaodan Zhu. Xiaodan Zhu participated in the
annotation and calculated the score of MMR.
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
94
a new task of building short summaries in response to a question. This task was carried on in
DUC 2004. In DUC 2005, the intention of modeling “real world complex question answering”
is more clear in the system task. It is to “synthesize from a set of 25-50 documents a brief, wellorganized, fluent answer to meet a need for information that cannot be met by just stating a
name, date, quantity, etc” [DUC, 2005]. However, according to our knowledge, summarization
techniques have not been explored by current QA systems. In our task, the information needed
is the clinical outcomes of an intervention on a disease, and we expect that summarization
techniques will help.
On the other hand, we also notice that multi-document summarization cannot replace QA.
One important difference between them, as pointed out in [Lin and Demner-Fushman, 2005], is
that summaries are compressible in length, i.e., summaries can contain various levels of details,
while answers are not. It is difficult to fix the length of answers.
Because of the difference between multi-document summarization and QA, we are not
taking the former as the full solution even in the answer generation of a QA task. Instead,
we expect that some multi-document summarization techniques can be adapted to the answer
generation module in some non-factoid QA tasks. In this chapter, we explore summarization
techniques to identify important pieces of information for answer construction.
6.2
Clinical Evidence as a benchmark
Evaluation of a multi-document summarization system is difficult, especially in the medical
domain where there is no standard annotated corpora available. However, we observe that
Clinical Evidence (CE) provides a benchmark to evaluate our work against. As mentioned in
Section 1.5.1, CE is a publication that reviews and consolidates experimental results for clinical
problems; it is updated every six months. Each section in CE covers a particular clinical
problem, and is divided into several subsections that summarize the evidence concerning a
particular medication (or a class of medications) for the problem, including results of clinical
trials on the benefits and harms of the medications. The information sources that CE draws
on include medical journal abstracts, review articles, and textbooks. Human experts read the
collected information and summarize it to get concise evidence on every specific topic. This is
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
95
the process of multi-document summarization. Thus, each subsection of CE can be regarded
as a human-written multi-document summary of the literature that it cites.
Moreover, we observed that, generally speaking, the summaries in CE are close to being
extracts (as opposed to rewritten abstracts). A citation for each piece of evidence is given explicitly, and it is usually possible to identify the original Medline abstract sentence upon which
each sentence of the CE summary is based. Therefore, we were able to create a benchmark for
our system by converting the summaries in CE into their corresponding extracted summary.
That is, we matched each sentence in the CE summary to the sentence in the Medline abstract
on which it was based (if any) by finding the sentence that contained most of the same key
concepts mentioned in the CE sentence (this is similar to Goldstein et al. [1999]).
Using CE in our work has an additional advantage. As new results of clinical trials are
published fairly quickly, we need to provide the latest information to clinicians. We hope that
this work will contribute to semi-automatic construction of summaries for CE.
6.3
Identifying important sentences
6.3.1 Method
We perform summarization at the sentence level, i.e., we extract important sentences from a
set of documents to form a summary. For this, we explore a supervised approach. Again, we
treat the problem as a classification task, determining whether a sentence is important or not.
The same SVM package as in Section 5.3.2 (parameters were set in the same manner) is taken
as our machine learning system.
In the classification, each sentence is assigned an importance value by the classifier (SVM) 2
according to a predefined set of features. Sentences with higher values are more important and
will be extracted to form a summary of the original documents. Different lengths of summaries
(at different compression ratios) are obtained by selecting different numbers of sentences according to their rank in the output of SVMs. In the summaries, the sequence of sentences is
2
SVM output for each data point in the test set is a signed distance (positive= important) from the separating
hyperplane. A higher value means that a sentence is more important.
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
96
kept the same as in their original documents.
6.3.2 Features to identify important sentences
We use the presence and polarity of an outcome, both manually annotated and determined by
the method described in the previous chapter, as features to identify important sentences. In
addition, we consider a number of other features that have been shown to be effective in text
summarization tasks:
Position of a sentence in an abstract Sentences near the start or end of a text are more likely
to be important. We experimented with three different ways of representing sentence position:
1. Absolute position: sentence i receives the value i − 1.
2. The value for sentence i is i/length-of-the-document (in sentence).
3. A sentence receives a value of 1 if it is at the beginning (first 10%) of a document, a
value of 3 if it is at the end (last 10%) of a document, a value of 2 if it is in between.
Sentence length A score reflecting the number of words in a sentence, normalized by the
length of the longest sentence in the document [Lin, 1999].
Numerical value A sentence containing numerical values may be more specific and therefore
more likely to be important. We tried three options for this feature:
1. Whether or not the sentence contains a numerical value (binary).
2. The number of numerical values in the sentence.
3. Whether or not the sentence contains the symbol ‘%’ (binary).
Maximum Marginal Relevancy (MMR)
MMR is a measure of “relevant novelty”, and it
is formulated using terminologies in information retrieval. Its aim is to find a good balance
between relevancy and redundancy. The hypothesis is that information is important if it is both
relevant to the topic of interest and least similar to previously selected information, i.e., its
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
97
marginal relevance is high. MMR is defined as a linear combination of a relevance measure
and a novelty measure [Carbonell and Goldstein, 1998]:
M M R = argmax[λ(Sim1 (Di , Q) − (1 − λ) max Sim2 (Di , Dj ))]
Di ∈R\S
Dj ∈S
where R is the ranked list of the retrieved documents; Q is a query; S contains a set of documents that have been selected from R, therefore, S is a subset of R; R\S is the set of documents
in R that have not been selected; D is a document; Sim1 (Di , Q) is the similarity between the
document Di and query Q; Sim2 (Di , Dj ) is the similarity between two documents; Sim1 can
be the same as Sim2 . Parameter λ controls the impact of novelty and redundancy in summarization.
We adapt the original definition of MMR to our problem. In our task, R and Q are the
same—the list of sentences in all relevant documents from which a summary will be constructed. Because we do not have a specific query set, we set Q to be the same as R, which
is often the case in multi-document summarization systems. S is the subset of sentences in R
already selected; R \ S is the set difference, i.e., the set of sentences in R that are not selected
so far; Sim1 is a similarity metric; and Sim2 is the same as Sim1 . According to the definition
of MMR, when λ = 1, no redundancy is considered in ranking the sentences, i.e., no sentence
will be excluded from the summary because it contains redundant information. When λ = 0,
diversity dominates the constructed summary.
In our experiments, to calculate Sim(Di , Q), the sentence Di and the set of documents
Q are represented by vectors of tf · idf values ((1 + tf ) × idf ) of the terms they contain.
The similarity is measured by the cosine distance between two vectors. Similarly, we can
calculate Sim(Di , Dj ). The score of marginal relevance of a sentence is used as a feature in
the experiment (referred to as feature MMR).
6.4
Data Set
The data set in this experiment is the same as in 5.3.2; 197 Medline abstracts cited in 24
subsections (summaries) in CE are used. The average compression ratio of the 24 summaries
in CE is 0.25. Out of the total 2298 abstract sentences, 784 contain a clinical outcome (34.1%).
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
98
The total number of sentences in the 24 summaries is 546, of which 295 sentences contain a
clinical outcome (54.0%). The percentage of sentences containing a clinical outcome in the
summaries is larger than in the original Medline abstracts, which matches our intuition that
sentences containing clinical outcomes are important.
6.5
Evaluation
In our experiment, we randomly select Medline abstracts that correspond to 21 summaries in
CE as the training set, and use the rest of the abstracts (corresponding to 3 summaries in CE)
as the test set. The results reported are the average of 50 runs. As the purpose is to observe
the behavior of different feature sets, the experimental process can be viewed as a glass box.
The system was evaluated by two methods: sentence-level evaluation and ROUGE, an n-grambased evaluation approach. Both of the two methods are commonly used in the summarization
community. Randomly selected sentences are taken as baseline summaries.
To evaluate the performance of features, the subsections in CE are viewed as ideal summaries of the abstracts that they cite. The corresponding extraction summaries are used in the
sentence-level evaluation, and the original CE summaries are used for ROUGE evaluation.
6.5.1 Sentence-level evaluation
In the experiment, we first observe the performance of using every single feature in the classification. Then, we combine different features and investigate the contribution of the information
on clinical outcomes and their polarity in this task.
Comparison of individual features
The precision and recall curves of summaries derived by using every single feature at different
compression ratios are plotted in Figure 6.1.
In the figure, the solid horizontal line shows the purely chance performance, which is the
baseline. The baseline has a precision of 0.25 because the average compression ratio of CE
summaries is 0.25, and the recall at different compression ratios is calculated accordingly.
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
99
0.5
pos1
pos2
pos3
num1
num2
num3
len
MMR
pol(auto)
pol(manual)
outcome(auto)
outcome(manual)
0.45
Precision
0.4
0.35
0.3
0.25
0.2
0
0.1
0.2
0.3
0.4
0.5
Recall
0.6
0.7
0.8
0.9
1
Figure 6.1: Comparison of features
The other four colored solid lines represent the performance of manually or automatically
identified clinical outcome and polarity. Although compression ratio is not shown explicitly in
the figure, lower compression ratios correspond to lower recall, and compression ratio of 0.25
approximately corresponds to recall of 0.38 on the curve of MMR in the figure. Therefore, the
left-hand part of the figure (recall less than 0.5) is more meaningful. Thus, our analysis will
focus on this part of the figure.
On the left-hand part of the figure, all four features of the presence and polarity of clinical outcomes are superior to the baseline performance. It is clear that knowledge about clinical
outcomes helps in this task. We can see that manually obtained information on presence of outcome outperforms MMR. When compression ratio is relatively low, manually obtained polarity
information also performs better than MMR.
For MMR measurement, different values of λ are tested. Higher values produce better
summaries, which indicates that relevance is much more important than novelty in this task.
The figure shows the best results of MMR (λ = 0.9). Not surprisingly, MMR is quite effective
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
100
Table 6.1: F-score of the summarization with different feature sets in sentence-level evaluation
Compression Ratio
0.1
0.2
0.25
0.3
0.4
Random
.15
.22
.25
.27
.30
MMR
.27
.34
.37
.39
.42
(1)
.26
.36
.40
.42
.44
(2)
.27
.38
.41
.43
.46
(3)
.29
.40
.44
.46
.48
(4)
.27
.38
.40
.43
.46
(5)
.31
.41
.44
.46
.48
(1): MMR+position+numerical value+length
(2): (1)+automatically identified polarity of clinical outcomes
(3): (1)+manually identified polarity of clinical outcomes
(4): (1)+automatically identified presence of clinical outcomes
(5): (1)+manually identified presence of clinical outcomes
in the task. Other features such as length and numerical value (option 1) also have good effects
on the performance.
Combining the features
When features are combined, some of their effects will be additive, and some will cancel out.
Table 6.1 shows the F-score of using different combinations of features at different compression
ratios. The results of MMR, position, and numerical features listed in the table are the best
results obtained (MMR (λ = 0.9), position (option 2), and numerical features (option 1)).
These results are compared to the results of adding information on presence or polarity of
clinical outcomes.
As shown in the table, on top of the combination of MMR, position, numerical value, and
sentence length features, manually annotated outcome information (either presence or polarity)
results in 3 to 5 points of improvement in the F-score. This clearly indicates the importance of
the presence/polarity information for the task. Nevertheless, automatically identified informa-
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
101
tion regarding presence/polarity improves the performance only slightly when combined with
all other features. This suggests the need for developing more accurate techniques for outcome
and polarity detection.
The additional benefit from further determining the polarity of the outcome rather than just
detecting the presence of an outcome is rather small. Intuitively, we had expected polarity
to provide more information on contradicting clinical outcomes and thus to help improve the
performance. Looking closely at the data, however, we found that one aspect accounting for
the observed result could be that although some sentences are different in polarity, they do not
form a contradiction. Rather, they may describe different clinical outcomes. Since some of the
outcomes are not important, they are not included in the summaries:
(35) H pylori eradication is highly effective in promoting ulcer healing and preventing subsequent ulcer recurrence.
(36) However, while ulcer symptoms are infrequent during follow-up, a proportion of patients appear to develop gastrooesophageal reflux after eradication.
Example (35) expresses a positive outcome and it is included in the summary. Example (36)
describes a negative outcome. Although the polarity in the two sentences is different, they do
not contradict each other. It turns out that the second outcome is not important enough to be
included in the summary.
In some cases, although the two different outcomes are contradictions, they may not be
included in the summary because one or both are not strong evidence, as shown in the following
examples:
(37) Results of North American studies of highest methodological quality confirm that H
pylori eradication markedly decreases ulcer recurrence.
(38) Nevertheless, 20% of patients in these studies had ulcer recurrence within 6 months,
despite successful cure of infection and no reported use of NSAIDs.
Example (38) shows negative evidence on ulcer recurrence, while example (37) shows positive
evidence on it. However, (37) is strong evidence while (38) is not. Hence, only the first
is included in the summary. As a result, further determining whether a clinical outcome is
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
102
Table 6.2: Comparison of feature sets in single summary at compression ratio 0.25
(1) : MMR+position+numerical value+length
> (1)
= (1)
< (1)
(1)+Polarity(auto)
10
7
7
(1)+Polarity(manual)
15
6
3
(1)+Outcome(auto)
11
6
7
(1)+Outcome(manual)
15
2
7
positive, negative, or neutral does not have provide more evidence to support the importance
of a sentence.
We also observed the performance of different combinations of features on every single
summary at compression ratio 0.25, and show the results of feature sets with and without presence or polarity information in Table 6.2. The numbers in the table are numbers of summaries.
For example, the number in the second row, the first column is 10, which means that in 10
summaries the performance (F-score) of using feature set (1)+Polarity(auto) is better than using feature set (1). The complete result is plotted in Appendix J.
6.5.2 ROUGE
As an alternative evaluation, we use the ISI ROUGE package [Lin, 2004], which compares a
summary generated by a text summarization system with a benchmark summary by considering overlapping units such as n-grams, word sequences, and word pairs (word sequences and
word pairs allow gaps between words). Our evaluation was carried out with various ROUGE
parameters. Unlike the sentence-level evaluation, the results showed little difference in the
performance of different combinations of features. Table 6.3 shows the ROUGE-L score of
three feature sets. ROUGE-L is a measure of the longest common subsequence in the two
texts to be compared, where the longest common subsequence of two sequences X and Y is a
common subsequence with maximum length. For example, let X = [there are four cats],
Y1 = [I have four cats] and Y2 = [four cute cats are playing with a ball], then ROUGE-
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
103
Table 6.3: ROUGE-L score of different feature sets
Compression Ratio
0.1
0.2
P
R
F
P
MMR
.46
.18
.25
.40
(1)
.46
.18
.25
(2)
.46
.18
.25
R
0.3
F
0.4
P
R
F
P
R
F
.31 .33
.35
.40
.36
.30
.45
.35
.41
.31 .34
.35
.40
.36
.30
.45
.35
.41
.31 .34
.35
.39
.36
.30
.46
.35
(1): MMR+position+numerical value+length
(2): automatically identified presence of clinical outcomes
L(X, Y1 ) = ROUGE-L(X, Y2 ).
As the table shows, adding position, numerical value, and length features to MMR does
not improve the F-score compared to using MMR alone. Furthermore, adding automatically
identified presence of outcomes does not make any difference in F-score. One reason for the
result that different combinations of features perform almost the same in ROUGE evaluation
could be that it is difficult for an overlap-based metric to capture the difference if the content of
two sets is similar. For example, only a small difference might be measured by ROUGE when
comparing the inclusion of both a positive and a negative clinical outcome of an intervention
in treatment of a disease in the summary with the inclusion of only one of them.
6.6
Summary
In this chapter, we proposed an approach of identifying relevant pieces of information in answer construction for a specific task, i.e., answering questions about clinical outcomes. We
explored multi-document summarization techniques to rank sentences according to their importance in the context of constructing answers in QA. Our hypothesis is that information on
presence or polarity of clinical outcomes in a sentence would help identify answer components. Such information is used as features, together with some other features that have been
shown effective in previous summarization tasks. We investigated every single feature and
show that in general, the performance of the summarization system using outcome informa-
C HAPTER 6. S ENTENCE E XTRACTION USING O UTCOME P OLARITY
104
tion is superior to the baseline, especially when the compression ratio is relatively low (around
the average compression ratio of the CE summaries). This result demonstrates the effectiveness of such information. We also found that when features are combined, manually annotated
presence/polarity information improves the performance achieved by using all other features.
This further reveals the value of such information in the task. However, using automatically
detected presence/polarity information results in only a slight improvement. Thus, our next
step will be to build a more accurate outcome information detection system. Another finding
is that there is almost no difference between presence and polarity of clinical outcomes in the
task. Nevertheless, polarity information is mandatory for answering some questions, such as
questions about side effects of interventions.
In this experimental setting, the questions are about general clinical outcomes such as
“What is the effect of intervention A on disease B?”. For more specific questions as in Chapter 4, finer-grained analysis needs to be performed to find the answer, e.g., identifying cores
of semantic classes. For any case, text summarization techniques that address the problem of
finding relevant information and avoiding redundancy, e.g., MMR and various features, can be
used to identify answer components in NFQA.
Chapter 7
Conclusion
In the research that this dissertation presents, we have analyzed characteristics of NFQA. We
found that answers to non-factoid questions are usually more complex than named entities, and
multiple pieces of information are often needed to construct a complete answer. We proposed
a novel approach to address these characteristics. Important subtasks in different components
of the new approach were identified, and automatic methods were developed to solve the problems.
7.1
Summary of contributions
The contributions of this research have been presented in Section 1.6. We summarize them
here, adding emphasis on evaluation results.
The QA approach based on semantic class analysis We use frames to represent semantic
classes in scenarios and proposed an approach taking semantic class analysis as the organizing
principle to answer non-factoid questions. This approach contains four major components:
• Detecting semantic classes in questions and answer sources
• Identifying properties of semantic classes
• Question-answer matching: exploring properties of semantic classes to find relevant
pieces of information
105
C HAPTER 7. C ONCLUSION
106
• Constructing answers by merging or synthesizing relevant information using relations
between semantic classes
We investigated NFQA in the context of clinical question answering, and focused on three
semantic classes that correspond to roles in the commonly accepted PICO format of describing
clinical scenarios. The three classes are: the problem of the patient, the intervention used to
treat the problem, and the clinical outcome.
Extracting semantic classes and analyzing their relations We used rule-based approaches
to identify clinical outcomes and relations between instances of interventions in sentences. In
the combination approach of outcome identification, a set of cue words that signal the occurrence of an outcome were collected and classified according to their PoS tags. For each PoS
category, the syntactic components it suggests were summarized to derive rules of identifying boundaries of outcomes. This approach can potentially be applied to identify or extract
other semantic classes. We identified six common relationships between different instances
of interventions in a sentence and developed a cue-word-based approach to identify the relations automatically. These approaches were evaluated on a small data set annotated by human
judges.
Identifying cores of semantic classes We showed how cores of interventions, problems, and
outcomes in a sentence can be identified automatically by developing an approach exploring
semi-supervised learning techniques. The evaluation shows that each type of feature used in
the approach made contributions to the classification, and they are complementary in the task.
The structure of the data is followed by the semi-supervised model. Therefore, unlabeled data
is effectively exploited in the classification.
Detecting polarity of clinical outcomes We developed a method using a supervised learning
model to automatically detect polarity of clinical outcomes. The results show that context
features and category features significantly improved classification accuracy compared to using
unigrams alone. This method has stable performance on different sources of medical text. We
also identified a cause of the bottleneck of performance of supervised approach in polarity
C HAPTER 7. C ONCLUSION
107
detection, and showed evidence by the results of manual experiments.
Extracting components for answers We built explicit connection between text summarization and answer construction in NFQA, i.e., both of them need to identify important information and avoid redundancy. We constructed a summarization system that explored a supervised
classification model to extract important sentences for answer construction. We investigated
the role of presence and polarity of clinical outcomes in this task. The evaluation shows that
presence/polarity of clinical outcomes helps the summarization. However, accuracy of automatic approaches is not high enough to make a substantial improvement in the performance.
An additional advantage of the polarity information is that it is mandatory when answering
questions about benefits or harms of interventions.
Generalization of the approach of semantic class analysis Our approach analyzes semantic classes involved in scenarios to find answers to non-factual questions. The importance of
semantic classes has been shown both in theory of linguistics and by the success of IE systems. The properties of semantic classes that we identified for the medical domain apply to
other domains as well. A core is the essence of an instance of a semantic class; hence, instances of any semantic class have cores. Polarity is a property for any semantic class that can
be evaluated as good, bad, or neutral, although most of work on polarity analysis focuses on
opinions. In each subproblem of analysis of semantic classes discussed in the thesis, we use
general rule-based or machine learning approaches, although some domain-specific features
are incorporated in some tasks. Therefore, we expect that our approach can be used to answer
non-factual questions in other domains as well.
7.2
Future work
7.2.1 Extensions
Short-term future work includes overcoming some limitations or extending the current work.
• Refine rules of using adjectives to identify clinical outcomes. In both detecting the
occurrence of outcomes and determining the boundaries, the usage of adjective cue
C HAPTER 7. C ONCLUSION
108
words is harder to describe as compared to nouns and verbs. Finer-grained rules will
be able to handle these cases better.
• Extend rules of analyzing relations between instances of interventions to deal with
ambiguity. Although ambiguous cue words found in the relation analysis indicate one
relation more often than the other, it will be helpful to have more specific rules that
disambiguate the occurrence of such a cue.
• Investigate approaches to automatically collect words that express more or less, good
or bad to generate the change phrase features in detecting polarity of clinical outcomes.
• Explore more sophisticated techniques to handle negations in polarity detection. Negation is a very complex linguistics phenomenon. In-depth analysis of negations, for example, categorizing expressions of negations and investigating the syntactic relations
of the expressions, can be performed to reduce errors.
• Perform feature selection to remove redundant features in the polarity detection task.
• Improve boundary detection of the core identification task. In our current work, detecting boundaries of cores uses the UMLS knowledge base. Since such a knowledge
base does not exist for some domains, machine learning approaches can be explored
to improve boundary detection by investigating syntactic constituents of cores.
• Use cores of semantic classes to identify answer components. Various strategies can
be taken in calculating the similarity of a question and a piece of information in the
answer source. For example, only cores may be taken into account, or the complete
description may be considered while giving more weights to cores.
• Explore unsupervised summarization techniques to extract answer components. Unsupervised methods are important when annotated data is not available.
7.2.2 Directions for future work
On the basis of our research in non-factoid QA, we envision several major directions for future
work.
C HAPTER 7. C ONCLUSION
109
Improving polarity classification Polarity is a property of many things, such as clinical
outcomes and opinions. Automatic polarity classification, as shown in this thesis and in the research on sentiment analysis, is a crucial issue in many applications including QA. To improve
polarity classification, we found three research problems of great interest:
• The granularity of polarity analysis. As described in Section 5.1, current polarity research is at three levels: words, sentences, and documents. Determining polarity of
words will contribute to higher-level analysis. The advantage of performing sentencelevel analysis is that a sentence often expresses a polarity that is rather complete and
independent. Nevertheless, the difficulty at this level is that a sentence may contain
more than one polarity unit and they may have different polarity. In addition, the polarity of a sentence may be related to its adjacent sentences. For example, a sentence
may continue the same positive/negative polarity as its previous sentence although its
own expression is neutral. Document-level analysis can be too coarse-grained. Take
the main research at the document-level, movie reviews, as an example. A review usually contains discussions on both positive and negative aspects of a movie. Although
the polarity of the whole document may show the general view of a reviewer, it will be
much more informative to identify what in the movie is successful and what is not in
the reviewer’s opinion. This will rely on sentence-level analysis. Therefore, we believe
it is important to investigate the difficulty of sentence-level analysis mentioned above,
and develop new approaches to address it.
• Feature selection in polarity classification. As discussed in Section 5.4, polarity of a
clinical outcome is often determined by a sequence of words in a sentence. It will be
interesting to see if the same observation can be obtained from other polarity detection tasks, such as sentence-level sentiment classification. Current work in the opinion
domain often uses the semantic orientation of a set of words to detect polarity. This
may indicate the existence of a set of words that fully determine the polarity. This
observation may also hold at the document level. Because the goal of document-level
sentiment classification is to get the general view of the reviewer, it may not be necessary to consider all words or polarity expressions in the document, as it is in current
C HAPTER 7. C ONCLUSION
110
work. Instead, the polarity of the document may be set by one or a small amount of
sentences.
• The role of in-depth semantic analysis in polarity classification. As pointed out in Pang
et al. [2002], the task of sentiment classification is more difficult than the traditional
topic categorization task. Some examples in the paper show that in sentiment classification, even if a sentence or a document contains a lot of positive expressions, it
may turn out to be negative. This suggests that semantics of languages play a more
important role in the polarity classification task compared to the topic categorization
task. However, dominant features in current approaches for polarity classification are
similar as in topic categorization, i.e., using bag-of-words. Therefore, an important research direction is to effectively incorporate appropriate features capturing semantics
in text, such as negations.
Using topic detection to find answer components In Chapter 6, we described the approach
of using multi-document summarization techniques to identify relevant pieces of information
in a set of relevant documents to construct the answer. It will be interesting to further investigate this approach in the case that a relevant document discusses more diverse topics. For
such documents, it may be useful to first identify which part in a document is about the topic
addressed by the question, and then to apply summarization techniques to extract important
information from that particular part. This will require detecting topics in a document.
The problem of topic detection has been investigated for text summarization [Nomoto and
Matsumoto, 2001, 2003; Hardy et al., 2002]. The idea is to detect major topics discussed in
a document, and then extract important information from each topic to compose a summary.
This approach addresses the importance and redundancy by selecting salient information from
diverse topics. Another related research area is topic segmentation [Caillet et al., 2004], which
recently is explored mostly in speech, dialogue, and news [Purver et al., 2006; Malioutov and
Barzilay, 2006; Rosenberg and Hirschberg, 2006].
We will use topic areas (a topic area is a segment of text discussing one topic) to find the
answer to a given question. Therefore, after topic areas in a document are identified, we need a
similarity measure to select the topic that is most similar to the question. Then, we will extract
C HAPTER 7. C ONCLUSION
111
important information from this topic area. In QA, we usually have a set of documents that are
relevant to the question. Hence, this process will be conducted for every document, and the
answer will be constructed on the basis of the information extracted from each document.
Exploring textual context of frames In our frame-based semantic class analysis approach,
an answer frame in a document will be matched to the question frame by comparing the corresponding semantic classes in the two frames. Then, information in the answer frame will
be extracted from the document to construct the answer. However, it may not be appropriate
to only extract the answer frame as it is usually not isolated from its textual context in the
document. Instead, this context often describes related information such as the background,
the preconditions, and the explanation of the answer frame. Missing this information can result in misleading answers. Therefore, another research direction of locating relevant pieces of
information as answer components is to explore discourse analysis.
One of the most popular discourse theories is the Rhetorical Structure Theory proposed by
Mann and Thompson [1986]. The central notion in the theory is that of rhetorical relation,
which is a relation that holds between two non-overlapping text spans. Some examples of
rhetorical relations are: justification, equivalence, contrast, cause, condition, and change-topic.
The text spans can be text sequences within a sentence [Marcu and Echihabi, 2002], or they
can be sentences [Saito et al., 2006].
We can perform discourse analysis to understand how sentences in the context of an answer frame are related to each other, i.e., if the sentences are connected by some important
rhetorical relations. The set of rhetorical relations that are important in this task can be collected manually or via some supervised learning approaches. To construct the answer, both the
answer frame and its important context determined by the rhetorical relations will be extracted
from the original document.
Cross-sentence scenario analysis Currently, our work is at the sentence level, i.e., slot fillers
of a frame are extracted from a single sentence. Consequently, if a sentence only contains the
instances of some semantic classes in a frame, the frame extracted from the sentence will have
some empty slots. Because empty slots may cause confusion in the matching process, we need
to resolve them by locating the correct instances in the document.
C HAPTER 7. C ONCLUSION
112
We identified two causes of empty slots. In one case, instances of a semantic class are
simply omitted in a sentence because they are described in the previous sentence(s). In such
case, the meaning of this sentence is easy to interpret by taking into account the previous
sentence(s). Hence, it is easy to fill the empty slots by examining the corresponding semantic
classes in the previous sentence(s). In the other case, in a sentence, some instances of semantic
classes are referred to by various expressions, e.g., pronouns. One way to address this case is
to explore coreference resolution to locate the correct slot fillers.
Coreference resolution is a major research area in computational linguistics. Coreference
can be tackled by exploring syntactic constraints, semantics, and discourse information. Recently, statistical approaches using various models have been developed to improve the accuracy of coreference resolution [Bergsma and Lin, 2006; Yang et al., 2006]. We need to examine current approaches in detail for possible solutions to our problem of resolving empty slots.
This problem, however, may present new features that require new approaches to coreference
resolution.
Appendix A
List of abbreviations
CE
Clinical Evidence
DUC
Document Understanding Conferences
EBM
Evidence-based Medicine
EBOC
Evidence-based On Call
EPoCare Evidence at the Point of Care
FBQA
Fact-based Question Answering
HITIQA
High-Quality Interactive Question Answering
IE
Information Extraction
IR
Information Retrieval
LLE
Locally Linear Embedding
MMR
Maximum Marginal Relevancy
MPQA
Multi-Perspective Question Answering
MRR
Mean Reciprocal Rank
MUC
Message Understanding Conference
NE
Named Entities
NFQA
Non-factoid Question Answering
QA
Question Answering
RBF
Radial Basis Function
SO
Semantic Orientation
SVM
Support-Vector Machine
UMLS
Unified Medical Language System
113
Appendix B
A subset of syntactic tags in Apple Pie
Parser
PoS tags:
CC: Coordinating conjunction; DT: Determiner; IN: Preposition or subordinating conjunction;
JJ: Adjective; JJR: Adjective, comparative; NN: Noun, singular or mass; NNP: Proper noun,
singular; NNS: Noun, plural; NNPS: Proper noun, plural; NNPX: NNP + NNPS; RB: Adverb;
TO: to; VB: Verb, base form; VBN: Verb, past participle; VBZ: Verb, 3rd person singular
present
Phrase tags:
ADVP: Adverb phrase; NP: Noun phrase; NPL: NP which has no NP in its decendents, lowest
NP; PP: Prepositional phrase; VP: Verb phrase; S: Sentence
114
Appendix C
The effect of σ in the RBF kernel in core
identification
This appendix shows the classification results using different σ values in the RBF kernel (cosine
distance is used here). Figure C.1 shows that when σ is in a reasonable range, the performance
of the classification is not sensitive to its change.
Accuracy of classification with different Sigma values
0.75
0.7
0.65
Accuracy
0.6
0.55
0.5
0.45
labeled data: 10%
labeled data: 30%
labeled data: 60%
0.4
0.35
−1
−0.5
0
0.5
1
1.5
2
2.5
3
3.5
Sigma (with a base 10 logarithmic scale)
Figure C.1: Classification results with different values for σ
115
4
Appendix D
Algorithm for Boundary Detection of
Clinical Outcomes
Input:
S = {S1 , · · · Sn }: a set of parsed sentences.
Output:
O = {O1 , · · · On }: clinical outcomes identified from S.
Notations:[lP , rP ] — the boundary of a word sequence P .
N P (w) — the innermost noun phrase containing the word w.
AP (w) — the innermost adjective phrase containing w.
C(w) — the clause containing w.
com([l1 , r1 ], [l2 , r2 ]) — combination of intersecting intervals;
return [min(l1 , l2 ), max(r1 , r2 )].
max([l1 , r1 ], [l2 , r2 ]) — maximum cover; return the larger interval.
1: For each sentence Si in S:
2:
Initialize the boundary set B; set the position counter pc to the start of S i .
3:
While pc does not reach the end of Si :
4:
get next word w.
5:
compute the extraction boundary for the following cases of w:
6:
(i) w ∈ [difference, superior, effective]:
7:
(a) identify N P (w).
8:
(b) identify the propositional phrase P P which immediately
follows N P (w).
9:
(c) record the extraction boundary [lN P (w) , rP P ] in B.
116
A PPENDIX D. A LGORITHM FOR B OUNDARY D ETECTION OF C LINICAL O UTCOMES 117
10:
(ii) w is a cue verb:
11:
(a) check negation of w and label the negative status on w.
12:
(b1) if w is active, then
identify E which is the noun phrase immediately following w.
13:
(b2) if w is passive, then
identify E which is the word sequence between the start of C(w) and w.
14:
15:
(c) record the boundary [lE , rE ] in B.
(iii) w is a cue noun:
16:
(a) identify N P (w).
17:
(b) identify the phrase P̂ which contains w and is one level higher
than N P (w).
18:
(c) if P̂ is a noun phrase, then record [lP̂ , rP̂ ] in B;
else, record [lN P (w) , rN P (w) ] in B.
19:
(iv) w is an adjective:
20:
(a) identify N P (w).
21:
(b) identify AP (w).
22:
(c) record max([lN P (w) , rN P (w) ], [lAP (w) , rAP (w) ]) in B,
23:
EndOfWhileLoop.
24:
Combine boundaries in B using combine function until there are no intersecting boundaries.
25:
Output outcome Oi – word sequences extracted from Si as indicated by the boundaries in B.
26: EndOfForLoop.
Appendix E
Sample output of MetaMap
Sentence:
It found that the combined rate of myocardial infarction, stroke, or death was slightly lower in
the lower dose than in the higher dose groups at 3 months.
Output of MetaMap:
Phrase: “It”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “found”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “that”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “the”
Meta Candidates (0): <none>
Meta Mappings: <none>
118
A PPENDIX E. S AMPLE OUTPUT OF M ETA M AP
119
Phrase: “combined”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “rate”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “of myocardial infarction”
Meta Candidates (6)
1000 Myocardial Infarction [Disease or Syndrome]
861 Infarction [Finding,Pathologic Function]
861 Myocardial [Functional Concept]
805 MI <2> (Without) [Qualitative Concept]
789 MIS (Mullerian duct inhibiting substance) [Amino Acid, Peptide, or Protein,Hormone]
789 Myocardium [Tissue]
Meta Mapping (1000)
1000 Myocardial Infarction [Disease or Syndrome]
Phrase: “stroke”
Meta Candidates (1)
1000 Stroke (Cerebrovascular accident) [Disease or Syndrome]
Meta Mapping (1000)
1000 Stroke (Cerebrovascular accident) [Disease or Syndrome]
Phrase: “or”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “death”
A PPENDIX E. S AMPLE OUTPUT OF M ETA M AP
Meta Candidates (3)
1000 Death <1> (Cessation of life) [Finding,Organism Function]
916 LIQUEFACTION [Laboratory or Test Result]
900 Expired [Functional Concept]
Meta Mapping (1000)
1000 Death <1> (Cessation of life) [Finding,Organism Function]
Phrase: “was”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “slightly lower”
Meta Candidates (4)
861 Lower [Functional Concept]
805 LO [Pharmacologic Substance]
694 SLIGHTLY [Idea or Concept]
623 Slight [Qualitative Concept]
Meta Mapping (888)
694 SLIGHTLY [Idea or Concept]
861 Lower [Functional Concept]
Phrase: “in the lower dose”
Meta Candidates (3)
789 dosage (Dosages) [Quantitative Concept]
694 Lower [Functional Concept]
638 LO [Pharmacologic Substance]
Meta Mapping (853)
694 Lower [Functional Concept]
789 dosage (Dosages) [Quantitative Concept]
120
A PPENDIX E. S AMPLE OUTPUT OF M ETA M AP
Phrase: “than”
Meta Candidates (0): <none>
Meta Mappings: <none>
Phrase: “in the higher dose groups”
Meta Candidates (5)
827 Groups [Intellectual Product]
701 High dose [Quantitative Concept]
627 High [Qualitative Concept]
627 High <1> (Euphoric mood) [Mental Process]
589 dosage (Dosages) [Quantitative Concept]
Meta Mapping (879)
701 High dose [Quantitative Concept]
827 Groups [Intellectual Product]
Phrase: “at 3 months”
Meta Candidates (1)
861 months (month) [Temporal Concept]
Meta Mapping (861)
861 months (month) [Temporal Concept]
121
Appendix F
Sample output of Minipar
This appendix shows the output of the Minipar parser on a sentence. The output of the parser
is used to construct the syntactic relations features in the core identification task.
122
A PPENDIX F. S AMPLE OUTPUT OF M INIPAR
Sentence:
Thrombolysis reduces the risk of dependency, but increases the chance of death.
Candidates of cores:
thrombolysis, dependency, death
Output of Minipar:
(
E0 (() fin C * )
1 (Thrombolysis N 2 s (gov reduce))
2 (reduces reduce V E0 i (gov fin))
E2 (() Thrombolysis N 2 subj (gov reduce) (antecedent 1))
3 (the Det 4 det (gov risk))
4 (risk N 2 obj (gov reduce))
5 (of Prep 4 mod (gov risk))
6 (dependency N 5 pcomp-n (gov of))
7 (, U 2 punc (gov reduce))
8 (but U 2 punc (gov reduce))
9 (increases increase V 2 conj (gov reduce))
E3 (() Thrombolysis N 9 subj (gov increase) (antecedent 1))
10 (the Det 11 det (gov chance))
11 (chance N 9 obj (gov increase))
12 (of Prep 11 mod (gov chance))
13 (death N 12 pcomp-n (gov of))
14 (. U * punc)
)
Relations:
(thrombolysis subj–of increase), (thrombolysis subj–of reduce)
(dependency pcomp-n–of of)
(death pcomp-n–of of)
Figure F.1: Example of dependency triples extracted from output of Minipar parser.
123
Appendix G
List of words for building CHANGE
PHRASES
features
This appendix shows the list of good, bad, more, and less words collected from CE in detecting
polarity of clinical outcomes.
Good:
cured
vitality
relaxing
benefit
tolerability
improvement
right
effective
stable
best
better
pleasurable
relaxation
favour
beneficial
safety
prevents
successful
satisfaction
significant
superior
contributions
reliability
robust
tolerated
improving
survival
favourable
reliable
recovered
judiciously
consciousness efficacy
prevented
satisfied
prevent
advantage
encouraging
tolerance
success
significance improved
improves
improve
improvements
Bad:
depression
acute
sore
outpatient
disabling
diabetes
difficulties
dysfunction
distorted
poorer
unable
prolonged
irritation
disruptive
pathological
mutations
disease
infection
124
A PPENDIX G. L IST OF WORDS FOR BUILDING CHANGE PHRASES FEATURES
125
harms
difficulty
weakened
inactive
stressors
hypertension
adverse
insomnia
relapsing
malignant
suffer
exacerbate
dryness
fever
overestimate
constipation
deposition
colic
tension
hazards
diarrhoea
weakness
irritability
insidious
distress
weak
cancer
emergency
risk
block
unsatisfactory
blinding
nausea
traumatic
wound
intention
loses
intensive
relapse
recurrent
extension
die
cancers
malaise
crying
toxic
injury
confounding
complaints
misuse
insignificant
poisoning
anoxic
amputation
death
nightmares
deteriorate
fatal
injuries
fatigue
invasive
suicide
chronic
relapsed
disturbances
confusion
died
fluctuating
severities
delusions
compulsions
conflict
trauma
cried
impair
severe
tremor
weaker
illness
inpatients
worry
rebound
worse
reversible
dizziness
attacks
pointless
disorders
dyskinesia
risks
fatty
negative
conflicting
upset
fishy
hard
harm
bleeding
inflammatory
hampered
underpowered
obstruction
headache
problem
bleeds
panic
loss
odds
retardation
dysfunctional
render
difficult
drowsiness
lack
suicidal
obsessions
impaired
cough
severity
suffering
violent
strokes
virus
stroke
flatulence
fibrates
blind
burning
faintness
suffered
threatening
misdiagnosing
bitter
excessive
diabetics
malfunction
abnormal
deterioration
bad
confounded
sadness
mortality
disturbance
agitated
attack
infections
negativistic
deaths
poor
wrong
worsening
adversely
insufficient
scarring
headaches
disability
overdose
serious
delayed
discomfort
sweating
morbidity
nerve
parkinson
toxicity
nervous
pain
stress
A PPENDIX G. L IST OF WORDS FOR BUILDING CHANGE PHRASES FEATURES
weakens
incorrect
disorder
worsened
malformations
blinded
rigidity
prolong
adversity
abuse
lacked
dyspepsia
sads
onset
failure
inadequate sensitivity
dementia
harmful
elevation
higher
126
impairment
Increase:
increase
enhance
exceed
enhancement
peaked
more
excess
Decrease:
below
fewer
lower
decrease
fall
low
reduce
decline
less
little
mild
drop
Appendix H
Using an RBF kernel in detecting polarity
of clinical outcomes
This appendix shows the accuracy of using a non-linear kernel, RBF (exp(−d 2 (xi , xj )/σ 2 ), in
SVMs in the task of detecting positive and negative clinical outcomes in CE text. Using the
feature set of presence of combining UNIGRAMS with DISEASE and MORE / LESS, the accuracy
of the classification obtained with different σ values is shown in Table H.1. We can see that
small σ stretches the distance between data points and makes the classification very difficult.
When σ is larger, the performance is not very sensitive to its change, and becomes relatively
stable.
Table H.1: Results of positive/negative classification using an RBF kernel
σ2
10−2
10−1
100
101
102
103
104
105
Accuracy (%)
34.9
34.9
34.9
87.9
91.9
92.3
92.7
92.7
127
Appendix I
Results of the summarization with
different feature sets in sentence-level
evaluation
Table I.1: Results of the summarization with different feature sets in sentence-level evaluation
Compression Ratio
0.1
0.2
0.25
0.3
0.4
P
R
F
P
R
F
P
R
F
P
R
F
P
R
F
Random
.25
.11
.15
.25
.20
.22
.25
.26
.25
.25
.31
.27
.25
.40
.30
MMR
.44
.19
.27
.38
.31
.34
.36
.38
.37
.36
.44
.39
.34
.57
.42
(1)
.44
.19
.26
.40
.33
.36
.40
.41
.40
.38
.48
.42
. 36
.58
.44
(2)
.45
.20
.27
.42
.35
.38
.40
.42
.41
.39
.49
.43
.37
.61
.46
(3)
.45
.20
.27
.41
.35
.38
.40
.42
.40
.39
.48
.43
.37
.61
.46
(4)
.49
.21
.29
.44
.38
.40
.43
.46
.44
.41
.52
.46
.38
.64
.48
(5)
.51
.22
.31
.45
.38
.41
.43
.46
.44
.42
.53
.46
.39
.65
.48
(1): MMR+position+numerical value+length
(2): (1)+automatically identified polarity of clinical outcomes
(3): (1)+manually identified polarity of clinical outcomes
(4): (1)+automatically identified presence of clinical outcomes
(5): (1)+manually identified presence of clinical outcomes
128
Appendix J
F-score of combinations of features in
each single summary
The F-score of each summary at compression ratio 0.25 is presented in Figure J.1 to observe
the performance of different combinations of features on every single summary. The diagram
at the top shows the results of features including presence of outcomes, and the one on the
bottom shows the results of features including polarity of outcomes.
129
A PPENDIX J. F- SCORE OF COMBINATIONS OF FEATURES IN EACH SINGLE SUMMARY 130
1
MMR
MMR+pos+num+len
MMR+pos+num+len+outcome(auto)
MMR+pos+num+len+outcome(manual)
F−Score
0.8
0.6
0.4
0.2
0
1
2
3
4
5
6
7
8
1
MMR
MMR+pos+num+len
MMR+pos+num+len+pol(auto)
MMR+pos+num+len+pol(manual)
0.8
F−Score
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Summary No.
0.6
0.4
0.2
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Summary No.
Figure J.1: The performance of different combinations of features in each summary
Bibliography
Alpha, S., Dixon, P., Liao, C., and Yang, C. (2001). Oracle at TREC 10: Filtering and question–
answering. In Proceedings of the Tenth Text Retrieval Conference (TREC 2001).
Ananiadou, S. and Tsujii, J., editors (2003). Proceedings of the ACL 2003 Workshop on Natural
Language Processing in Biomedicine. Association for Computational Linguistics (ACL),
PA, USA.
Aronson, A. R. (2001). Effective mapping of biomedical text to the UMLS metathesaurus:
the MetaMap program. In Proceedings of the American Medical Informatics Association
Symposium, pages 17–21.
Baker, C. F., Fillmore, C. J., and Cronin, B. (2003). The structure of the Framenet database.
International Journal of Lexicography, 16(3):281–296.
Ball, C. M. and Phillips, R. S. (2001). Evidence-based on Call: Acute Medicine. Harcourt
Publishers Limited, London.
Barbosa, D., Barta, A., Mendelzon, A., Mihaila, G., Rizzolo, F., and Rodriguez-Gianolli, P.
(2001). Tox — the toronto xml engine. In Proceedings of the International Workshop on
Information Integration on the Web, Rio de Janeiro, 2001.
Barton, S., editor (2002). Clinical Evidence. BMJ Publishing Group, London.
Berger, A., Caruana, R., Cohn, D., Freitag, D., and Mittal, V. (2000). Bridging the lexical
chasm: statistical approaches to answer-finding. In Proceedings of the 23th International
Conference on Research and Development in Information Retrieval (SIGIR- 2000), pages
192–199.
131
BIBLIOGRAPHY
132
Bergsma, S. and Lin, D. (2006). Bootstrapping path-based pronoun resolution. In Proceedings
of the 44th Annual Meeting of the Association for Computational Linguistics.
Blum, A. and Mitchell, T. (1998). Combining labeled and unlabeled data with co-training.
In Proceedings of the 11th Annual Conference on Computational Learning Theory, pages
92–100.
Brill, E. (1993). A Corpus-Based Approach to Language Learning (PhD thesis). University of
Pennsylvania, PA, USA.
Caillet, M., Pessiot, J. F., Amini, M. R., and Gallinari, P. (2004).
Unsupervised learn-
ing with term clustering for thematic segmentation of texts. In Proceedings of Recherche
d’Information Assistée par Ordinateur (RIAO).
Carbonell, J. and Goldstein, J. (1998). The use of MMR, diversity-based reranking for reordering documents and producing summaries. In Proceedings of the 21st annual international
ACM SIGIR conference on Research and development in information retrieval, pages 335–
336.
Cardie, C., Wiebe, J., Wilson, T., , and Litman, D. (2003). Combining low-level and summary
representations of opinions for multi-perspective question answering. In Proceedings of the
AAAI Spring Symposium: New Directions in Question Answering, pages 20–27.
Català, N., Castell, N., and Martı́n, M. (2003). A portable method for acquiring information
extraction patterns without annotated corpora. Natural Language Engineering, 9(2):151–
179.
Chang, C.-C. and Lin, C.-J. (2001). LIBSVM–a library for support vector machines. In
http://www.csie.ntu.edu.tw/ cjlin/libsvm/.
Cheng, G. Y. (2004). A study of clinical questions posed by hospital clinicians. J Med Libr
Assoc., 92(4):445–458.
Chieu, H. L. and Ng, H. T. (2003). Named entity recognition with a maximum entropy approach. In Proceedings of 7th Conference on Computational Natural Language Learning,
pages 160–163.
BIBLIOGRAPHY
133
Cimino, J. J. (1996). Linking patient information systems to bibliographic resources. Methods
of Information in Medicine, 35(2):122–126.
Clarke, C. L. A., Cormack, G. V., Kisman, D. I. E., and Lynam, T. R. (2000). Question
answering by passage selection. In Proceedings of the Ninth Text Retrieval Conference
(TREC 2000).
Clarke, C. L. A., Cormack, G. V., and Lynam, T. R. (2001). Exploiting redundancy in question answering. In Proceedings of the 24th International Conference on Research and Development in Information Retrieval, September 9-13, 2001, New Orleans, Louisiana, USA
(SIGIR-2001), pages 358–365.
Collins, M. and Singer, Y. (1999). Unsupervised models for named entity classification. In Proceedings of the 1999 Joint SIGDAT Conference on Empirical Methods in Natural Language
Processing and Very Large Corpora.
Cucerzan, S. and Yarowsky, D. (1999). Language independent named entity recognition combining morphological and contextual evidence. In Proceedings of the 1999 Joint SIGDAT
Conference on Empirical Methods in Natural Language Processing and Very Large Corpora.
Diekema, Yilmazel, A., Chen, O., Harwell, J., Sarah, H., L., , Liddy, and D., E. (2003). What
do you mean? finding answers to complex questions. In Proceedings of the AAAI Spring
Symposium: New Directions in Question Answering.
Dowty, D. R. (1991). Thematic proto-roles and argument selection. Language, 67(3):547–619.
DUC (2005). Document understanding conferences. In http://duc.nist.gov/duc2005.
Dumais, S., Banko, M., Brill, E., Lin, J., and Ng, A. (2002). Web question answering: is
more always better? In Proceedings of the 25th International Conference on Research and
Development in Information Retrieval, August 11-15, 2002, Tampere, Finland (SIGIR-2002),
pages 291–298.
BIBLIOGRAPHY
134
Ebell, M. H. (1999). Information at the point of care: answering clinical questions. J AM
Board Fam Pract., 12(3):225–235.
Ferret, O., Grau, B., Hurault-Plantet, M., Illouz, G., Monceaux, L., Robba, I., and Vilnat, A.
(2001). Finding an answer based on the recognition of the question focus. In Proceedings of
the Tenth Text Retrieval Conference, National Institute of Standards and Technology (NIST),
Nov. 13-16, Gaithersburg, Maryland, USA (TREC 2001), pages 362–370.
Fillmore, C. J. (1976). Frame semantics and the nature of language. Annals of the New York
Academy of Sciences: Conference on the Origin and Development of Language and Speech,
280:20–32.
Florian, R., Ittycheriah, A., Jing, H., and Zhang, T. (2003). Named entity recognition through
classifier combination. In Proceedings of 7th Conference on Computational Natural Language Learning, pages 168–171.
Gildea, D. and Jurafsky, D. (2002). Automatic labeling of semantic roles. Computational
Linguistics, 28(3):245–288.
Goldstein, J., Kantrowitz, M., Mittal, V., and Carbonell, J. (1999). Summarizing text documents: Sentence selection and evaluation metrics. In Proceedings of the 22st annual international ACM SIGIR conference on Research and development in information retrieval,
pages 121–128.
Gorman, P., Ash, J., and Wykoff, L. (1994). Can primary care physicians’ questions be
answered using the medical journal literature?
Bulletin of Medical Library Association,
82(2):140–146.
Harabagiu, S., Maiorano, S., Moschitti, A., and Bejan, C. (2004). Intentions, implicatures
and processing of complex questions. In Workshop on Pragmatics of Question Answering,
Human Language Technology Conference of the North American Chapter of the Association
for Computational Linguistics, pages 31–42.
Harabagiu, S. M., Moldovan, D., Paşca, M., Mihalcea, R., Surdeanu, M., Bunescu, R., Gı̂rju,
R., Rus, V., and Morǎrescu, P. (2001a). The role of lexico-semantic feedback in open-
BIBLIOGRAPHY
135
domain textual question–answering. In Proceedings of the 39th Meeting of the Association
for Computational Linguistics (ACL-2001), pages 274–281.
Harabagiu, S. M., Moldovan, D. I., Paşca, M., Surdeanu, M., Mihalcea, R., G îrju, R., Rus,
V., Lăcătusu, F., Morărescu, P., and Bunescu, R. C. (2001b). Answering complex, list and
context questions with lcc’s question-answering server. In Proceedings of the Tenth Text
Retrieval Conference (TREC 2001).
Harabagiu, S. M., Paşca, M. A., and Maiorano, S. J. (2000). Experiments with open-domain
textual question answering. In Proceedings of the 18th International Conference on Computational Linguistics (COLING-2000), pages 292–298.
Hardy, H., Shimizu, N., Strzalkowski, T., Ting, L., Zhang, X., and Wise, G. B. (2002). Crossdocument summarization by concept classification. In Proceedings of the 25th annual international ACM SIGIR conference on Research and development in information retrieval,
pages 121–128.
Hickl, A., Lehmann, J., Williams, J., and Harabagiu, S. (2004). Experiments with interactive
question answering in complex scenarios. In Workshop on Pragmatics of Question Answering, Human Language Technology Conference of the North American Chapter of the
Association for Computational Linguistics, pages 60–69.
Hovy, E., Gerber, L., Hermjakob, U., Junk, M., and Lin, C.-Y. (2000). Question answering in
Webclopedia. In Proceedings of the Ninth Text Retrieval Conference (TREC 2000).
Hovy, E., Hermjakob, U., and Lin, C.-Y. (2001). The use of external knowledge in factoid QA.
In Proceedings of the Tenth Text Retrieval Conference, National Institute of Standards and
Technology (NIST), Nov. 13-16, Gaithersburg, Maryland, USA (TREC 2001), pages 644–
652.
Huang, T.-M., Kecman, V., and Kopriva, I. (2006). Kernel Based Algorithms for Mining Huge
Data Sets. Springer, Berlin, Germany.
Ittycheriah, A., Franz, M., and Roukos, S. (2001). IBM’s statistical question answering system
- TREC-10. In Proceedings of the Tenth Text Retrieval Conference, National Institute of
BIBLIOGRAPHY
136
Standards and Technology (NIST), Nov. 13-16, Gaithersburg, Maryland, USA (TREC 2001),
pages 258–264.
Jaakkola, T., Diekhans, M., and Haussler, D. (1999). Using the fisher kernel method to detect
remote protein homologies. In Proceedings of the Seventh International Conference on
Intelligent Systems for Molecular Biology, pages 149–158.
Joachims, T. (2002). SVMlight support vector machine.
Klein, D., Smarr, J., Nguyen, H., and Manning, C. D. (2003). Named entity recognition with
character-level models. In Proceedings of 7th Conference on Computational Natural Language Learning, pages 180–183.
Kohomban, U. S. and Lee, W. S. (2005). Learning semantic classes for word sense disambiguation. In Proceedings of the 43rd Annual Meeting of the Association for Computational
Linguistics, pages 34–41.
Lee, G. G., Seo, J., Lee, S., Jung, H., Cho, B.-H., Lee, C., Kwak, B.-K., Cha, J., Kim, D.,
An, J., Kim, H., and Kim, K. (2001). SiteQ: engineering high performance QA system
using lexico-semantic pattern matching and shallow NLP. In Proceedings of the Ninth Text
Retrieval Conference (TREC 2000).
Lin, C.-Y. (1999). Training a selection function for extraction. In Proceedings of the Eighteenth Annual International ACM Conference on Information and Knowledge Management
(CIKM), pages 55–62.
Lin, C.-Y. (2004). ROUGE: a package for automatic evaluation of summaries. In Proceedings
of the 42nd Annual Meeting of the Association for Computational Linguistics, Workshop on
Text Summarization Branches Out.
Lin, D. (1994). Principar – an efficient, broad-coverage, principle-based parser. In Proceedings
of the 15th International Conference on Computational Linguistics, pages 482–488.
Lin, D. (1998). Automatic retrieval and clustering of similar words. In Proceedings of the 17th
International Conference on Computational Linguistics, pages 768 – 774.
BIBLIOGRAPHY
137
Lin, J. and Demner-Fushman, D. (2005). Evaluating summaries and answers: Two sides of the
same coin? In Proceedings of the ACL 2005 Workshop on Intrinsic and Extrinsic Evaluation
Measures for MT and/or Summarization.
Ma, J., Zhao, Y., Ahalt, S., and Eads, D. (2003). OSU SVM classifier Matlab toolbox. In
http://svm.sourceforge.net/docs/3.00/api/.
Malioutov, I. and Barzilay, R. (2006). Minimum cut model for spoken lecture segmentation.
In Proceedings of 44th Annual Meeting of the Association for Computational Linguistics.
Mann, W. and Thompson, S. (1986). Rhetorical Structure Theory: Description and Construction of Text Structures. Kluwer Academic Publishers, Boston, U.S.A.
Manning, C. and Sch´’utze, H. (1999). Foundations of Statistical Natural Language Processing.
MIT Press, Cambridge, MA, U.S.A.
Marcu, D. and Echihabi, A. (2002). An unsupervised approach to recognizing discourse relations. In Proceedings of the 40th Annual Meeting of the Association for Computational
Linguistics, pages 368–375.
Mendonça, E. A., Cimino, J. J., Johnson, S. B., and Seol., Y.-H. (2001). Accessing heterogeneous sources of evidence to answer clinical questions. Journal of Biomedical Informatics,
34:85–98.
Moldovan, D. and Harabagiu, S. M. (2000). The structure and performance of an open-domain
question answering system. In Proceedings of the 38th Meeting of the Association for Computational Linguistics, Oct. 1-8, 2000, Hong Kong, China (ACL-2000), pages 563–570.
Moldovan, D., Harabagiu, S. M., Paşca, M., Mihalcea, R., Goodrum, R., Gı̂rju, R., and Rus,
V. (1999). LASSO: A tool for surfing the answer net. In Proceedings of the Eighth Text
Retrieval Conference (TREC 1999).
MUC
(1995).
Message
Understanding
Conferences.
http://www.cs.nyu.edu/cs/faculty/grishman/muc6.html.
National Library of Medicine (2004). SPECIALIST parser. In SPECIALIST Text Tools.
In
BIBLIOGRAPHY
138
Niu, Y. and Hirst, G. (2004). Analysis of semantic classes in medical text for question answering. In Proceedings of the 42nd Annual Meeting of the Association for Computational
Linguistics, Workshop on Question Answering in Restricted Domains, pages 54–61.
Niu, Y., Hirst, G., McArthur, G., and Rodriguez-Gianolli, P. (2003). Answering clinical questions with role identification. In Proceedings of 41st Annual Meeting of the Association
for Computational Linguistics, Workshop on Natural Language Processing in Biomedicine,
pages 73–80.
Niu, Y., Zhu, X., and Hirst, G. (2006). Using outcome polarity in sentence extraction for medical question-answering. In Proceedings of the American Medical Informatics Association
2006 Annual Symposium, pages 599–603.
Niu, Y., Zhu, X., Li, J., and Hirst, G. (2005). Analysis of polarity information in medical text.
In Proceedings of the American Medical Informatics Association 2005 Annual Symposium,
pages 570–574.
Nomoto, T. and Matsumoto, Y. (2001). A new approach to unsupervised text summarization.
In Proceedings of the 24th annual international ACM SIGIR conference on Research and
development in information retrieval, pages 26–34.
Nomoto, T. and Matsumoto, Y. (2003). The diversity-based approach to open-domain text summarization. Information Processing and Management: an International Journal, 39(3):363–
389.
Paşca, M. and Harabagiu, S. M. (2001a). The informative role of wordnet in open-domain
question answering. In Proceedings of the NAACL 2001 Workshop on WordNet and Other
Lexical Resources: Applications, Extensions and Customizations, pages 138–143.
Paşca, M. A. and Harabagiu, S. M. (2001b). High performance question/answering. In Proceedings of the 24th International Conference on Research and Development in Information
Retrieval, September 9-13, 2001, New Orleans, Louisiana, USA (SIGIR-2001), pages 366–
374.
BIBLIOGRAPHY
139
Pang, B. and Lee, L. (2004). A sentimental education: sentiment analysis using subjectivity
smmarizaiton based on minimum cuts. In Proceedings of the 42th Annual Meeting of the
Association for Computational Linguistics, pages 271–278.
Pang, B., Lee, L., and Vaithyanathan, S. (2002). Thumbs up? sentiment classification using
machine learning techniques. In Proceedings of 2002 Conference on Empirical Methods in
Natural Language Processing (EMNLP), pages 79–86.
Prager, J., Brown, E., and Coden, A. (2000). Question–answering by predictive annotation. In
Proceedings of the 23th International Conference on Research and Development in Information Retrieval (SIGIR-2000), pages 184–191.
Purver, M., Körding, K., Griffiths, T., and Tenebaum, J. (2006). Unsupervised topic modelling
for multi-party spoken discourse. In Proceedings of 44th Annual Meeting of the Association
for Computational Linguistics.
Ray, S. and Craven, M. (2001). Representing sentence structure in hidden Markov models for
information extraction. In Proceeding of 17th International Joint Conferences on Artificial
Intelligence, pages 1273–1279.
Richardson, W. S., Wilson, M. C., Nishikawa, J., and Hayward, R. S. (1995). The well-built
clinical question: a key to evidence-based decisions. ACP J Club, 123(3):12–13.
Riloff, E. (1999). Information extraction as a stepping stone toward story understanding. Computational Models of Reading and Understanding.
Rosario, B. and Hearst, M. A. (2004). Classifying semantic relations in bioscience texts.
In Proceedings of 42nd Annual Meeting of the Association for Computational Linguistics,
pages 431–438.
Rosenberg, A. and Hirschberg, J. (2006). Story segmentation of broadcast news in english,
mandarin, and arabic. In Proceedings of the Human Language Technology Conference of
the NAACL, Companion Volume: Short Papers, pages 125–128.
BIBLIOGRAPHY
140
Roweis, S. and Saul, L. (2000). Nonlinear dimensionality reduction by locally linear embedding. Science, 290(5500):2323–2326.
Sackett, D. L. and Straus, S. E. (1998). Finding and applying evidence during clinical rounds:
the “Evidence Cart”. Journal of the American Medical Association, 280(15):1336–1338.
Sackett, D. L., Straus, S. E., Richardson, W. S., Rosenberg, W., and Haynes, R. B. (2000).
Evidence-Based Medicine: How to Practice and Teach EBM. Harcourt Publishers Limited,
Edinburgh.
Saito, M., Yamamoto, K., and Sekine, S. (2006). Using phrasal patterns to identify discourse
relations. In Proceedings of the Human Language Technology Conference of the NAACL,
Companion Volume: Short Papers, pages 133–136.
Sang, E. F. T. K. and Meulder, F. D. (2003). Introduction to the conll-2003 shared task:
Language-independent named entity recognition. In Proceedings of Conference on Computational Natural Language Learning, pages 142–147.
Sekine, S. (1997). Apple pie parser. In http://nlp.cs.nyu.edu/app/.
Small, S., Strzalkowski, T., Liu, T., Ryan, S., Salkin, R., Shimizu, N., Kantor, P., Kelly, D.,
Rittman, R., Wacholder, N., , and Yamrom, B. (2004). Hitiqa: Scenario based question answering. In Workshop on Pragmatics of Question Answering, Human Language Technology
Conference of the North American Chapter of the Association for Computational Linguistics, pages 52–59.
Soricut, R. and Brill, E. (2006). Automatic question answering using the web: Beyond the
factoid. Information Retrieval – Special Issue on Web Information Retrieval, 9:191–206.
Soubbotin, M. M. (2001). Patterns of potential answer expressions as clues to the right answers.
In Proceedings of the Tenth Text Retrieval Conference, National Institute of Standards and
Technology (NIST), Nov. 13-16, Gaithersburg, Maryland, USA (TREC 2001), pages 293–
302.
BIBLIOGRAPHY
141
Srihari, R. and Li, W. (1999). Information extraction supported question answering. In Proceedings of the Eighth Text Retrieval Conference, National Institute of Standards and Technology (NIST), Nov. 17-19, Gaithersburg, Maryland, USA (TREC 1999), pages 185–196.
Stoyanov, V., Cardie, C., and Wiebe, J. (2005). Multi-perspective question answering using
the opqa corpus. In Proceedings of Human Language Technology Conference/Conference
on Empirical Methods in Natural Language Processing.
Straus, S. E. and Sackett, D. L. (1999). Bringing evidence to the point of care. Journal of the
American Medical Association, 281:1171–1172.
TREC (2001). Text retrieval conference. In http://trec.nist.gov/.
Turney, P. (2002). Thumbs up or thumbs down? semantic orientation applied to unsupervised
classification of reviews. In Proceedings of the 40th Annual Meeting of the Association for
Computational Linguistics, pages 417–424.
Valin, V. and Robert, D. (1993). A synposis of role and reference grammar. In Robert, D. and
Valin, V., editors, Advances in Role and Reference Grammar, pages 1 – 166. John Benjamins
Publishing Company, Amsterdam.
Voorhees, E. M. (2001). Overview of the TREC 2001 question answering track. In Proceedings
of the Tenth Text Retrieval Conference (TREC 2001).
Whitelaw, C., Garg, N., and Argamon, S. (2005). Using appraisal groups for sentiment analysis. In Proceedings of the 14th ACM international conference on Information and knowledge
management, pages 625–631.
Yang, X., Su, J., and Tan, C. L. (2006). Kernel-based pronoun resolution with structured
syntactic knowledge. In Proceedings of the 44th Annual Meeting of the Association for
Computational Linguistics.
Yu, H. and Hatzivassiloglou, V. (2003). Towards answering opinion questions: Separating facts
from opinions and identifying the polarity of opinion sentences. In Proceedings of the 2003
Conference on Empirical Methods in Natural Language Processing, pages 129–136.
BIBLIOGRAPHY
142
Zhu, X. and Ghahramani, Z. (2002). Learning from labeled and unlabeled data with label
propagation. In Carnegie Mellon University, Center for Automated Learning and Discovery,
Technical Report CMU-CALD-02-107.
Zhu, X., Ghahramani, Z., and Lafferty, J. (2003). Semi-supervised learning using Gaussian
fields and harmonic functions. In Proceedings of the 20th International Conference on Machine Learning.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement