Opinion search engine

Opinion search engine
US 20110078157A1
(19) United States
(12) Patent Application Publication (10) Pub. No.: US 2011/0078157 A1
Sun et al.
(54)
(43) Pub. Date:
OPINION SEARCH ENGINE
Publication Classi?cation
(51)
(75) Inventors:
Mar. 31, 2011
.
Int. Cl.
G06F 17/30
Jlan-Tao Sun, Beijing (CN);
(2006.01)
G06F 7/00
(200601)
iiaolghgan ?TCLI\IB)BijGiHg (CV15); Peng
(52) us. Cl. ...................................................... .. 707/749
Beijing (CN); Ke Tang, Beijing
(CN); Zheng Chen, Beijing (CN)
(57)
ABSTRACT
A computer-readable storage medium having stored thereon
u,
e1j1ng
;
ang
ang,
computer-executable instructions Which, When executed by a
computer, cause the computer to implement an opinion
(73) Assignee;
Microsoft Corporation’ Redmond’
search engine. The instructions to implement an opinion
WA (Us)
search engine cause the computer to collect opinion data
about one or more objects from the lntemet, extract metadata
about the opinion data from the opinion data, remove dupli
cate metadata from the metadata to generate a resulting meta
(21) Appl' N05
12/568,702
data, categorize the resulting metadata for similar objects
according to one or more taxonomies from one or more Web
sites on the lntemet and rank the similar objects based on the
(22) Filed:
Sep. 29, 2009
categorized metadata.
2 0
COLLECT OPINION DATA FROM
INTERNET
I
EXTRACT METADATA FROM OPINION
DATA
I
AGGREGATE METADATA
I
CATEGORIZE METADATA
I
RANK OPINION DATA
K
220
K
230
T
240
K
2
K
Patent Application Publication
Mar. 31, 2011 Sheet 1 0f 4
US 2011/0078157 Al
on
Patent Application Publication
Mar. 31, 2011 Sheet 2 0f 4
US 2011/0078157 A1
20
COLLECT OPINION DATA FROM
INTERNET
I
EXTRACT METADATA FROM OPINION
DATA
I
AGGREGATE METADATA
I
CATEGORIZE METADATA
I
RANK OPINION DATA
FIG. 2
K2
0
220
K
230
K
240
K
250
K
Patent Application Publication
Mar. 31, 2011 Sheet 3 0f 4
0
FIG. 3
US 2011/0078157 A1
Patent Application Publication
Mar. 31, 2011 Sheet 4 0f 4
US 2011/0078157 A1
MONTH1
K [MONTH 2
V
TOTAL
POS|T|VE
NEGATIVE
TOTAL
POSITIVE
+34%
NEGATIVE
-0.1%
-1%
0%
1%
FIG. 4
2%
3%
US 2011/0078157 A1
OPINION SEARCH ENGINE
BACKGROUND
[0001] Internet users (reviewers) use various Web-based
forums such as blogs and revieW Websites to express their
Mar. 31, 2011
essential features of the claimed subject matter, nor is it
intended to be used to limit the scope of the claimed subject
matter. Furthermore, the claimed subject matter is not limited
to implementations that solve any or all disadvantages noted
in any part of this disclosure.
opinion about any topic of interest such books, hotels, con
sumer products, political policy and the like. These opinions
about various topics of interest may be referred to as opinion
data. Opinion data is typically used to help consumers make
an informed purchase decision about an item that they may
Wish to purchase based on the opinions of other consumers.
Opinion data is also used to help companies learn more about
hoW their customers rate their products, customer sentiment
for their products and customer satisfaction for their products.
BRIEF DESCRIPTION OF THE DRAWINGS
[0005]
FIG. 1 illustrates a schematic diagram of a comput
ing system in Which the various techniques described herein
may be incorporated and practiced.
[0006]
FIG. 2 illustrates a How diagram of a method for
implementing an opinion search engine in accordance With
one or more implementations of various techniques described
net, opinion data is voluminous and includes a plethora of
herein.
[0007] FIG. 3 illustrates a graph of a trend of opinion data
for a single product in accordance With one or more imple
diverse topics, Which makes it dif?cult for users to locate
mentations of various techniques described herein.
Since anyone can add an opinion about anything on the Inter
relevant opinions related to their topic of interest.
[0008]
FIG. 4 illustrates graphs of opinion change data for
a single product in accordance With one or more implemen
SUMMARY
[0002]
Described herein are implementations of various
technologies for implementing an opinion search engine. In
tations of various techniques described herein.
DETAILED DESCRIPTION
one implementation, a computer application may access each
Webpage available on the Internet and determine Whether
[0009] In general, one or more implementations described
herein are directed to implementing an opinion search engine.
each Webpage contains opinion data. If the Webpage contains
opinion data, the computer application may store the
Various techniques for implementing an opinion search
Webpage on a ?rst database. The computer application may
then extract information pertaining to the opinion data from
each Webpage stored on the ?rst database. While extracting
the information pertaining to the opinion data, the computer
application may determine Whether the information corre
sponds to a particular category of information or an informa
tion data type. The computer application may then store the
information pertaining to the opinion data in a second data
base according to the information’s particular category. In
this manner, the second database may be structured according
to the properties of the information.
[0003] The computer application may then remove dupli
cative information and group similar information together. In
addition to removing duplicative information from the second
database and group similar information together, the com
puter application may determine a taxonomy for the informa
tion in the second database. The taxonomy may refer to a
classi?cation scheme such that the information in the second
database may be organiZed in a hierarchal structure. In order
to determine an appropriate taxonomy for all of the informa
tion in the second database, the computer application may
examine the taxonomies of each Webpage listed in the second
database. In one implementation, the computer application
may leverage each Web page’s link paths and site map infor
mation to determine the appropriate taxonomy for all of the
information in the second database. After determining the
appropriate taxonomies, the computer application may orga
niZe the information pertaining to the opinion data in the
second database according to the appropriate taxonomy. The
computer application may then use the organiZed information
in the second database to rank products, product categories,
opinions, and other opinion-related subject matter on the
Internet.
[0004]
The above referenced summary section is provided
to introduce a selection of concepts in a simpli?ed form that
are further described beloW in the detailed description sec
tion. The summary is not intended to identify key features or
engine Will be described in more detail With reference to
FIGS. 1-4.
[0010] Implementations of various technologies described
herein may be operational With numerous general purpose or
special purpose computing system environments or con?gu
rations. Examples of Well knoWn computing systems, envi
ronments, and/or con?gurations that may be suitable for use
With the various technologies described herein include, but
are not limited to, personal computers, server computers,
hand-held or laptop devices, multiprocessor systems, micro
processor-based systems, set top boxes, programmable con
sumer electronics, netWork PCs, minicomputers, mainframe
computers, distributed computing environments that include
any of the above systems or devices, and the like.
[0011] The various technologies described herein may be
implemented in the general context of computer-executable
instructions, such as program modules, being executed by a
computer. Generally, program modules include routines, pro
grams, objects, components, data structures, etc. that per
forms particular tasks or implement particular abstract data
types. The various technologies described herein may also be
implemented in distributed computing environments Where
tasks are performed by remote processing devices that are
linked through a communications netWork, e. g., by hardWired
links, Wireless links, or combinations thereof. In a distributed
computing environment, program modules may be located in
both local and remote computer storage media including
memory storage devices.
[0012] FIG. 1 illustrates a schematic diagram of a comput
ing system 100 in Which the various technologies described
herein may be incorporated and practiced. Although the com
puting system 100 may be a conventional desktop or a server
computer, as described above, other computer system con
?gurations may be used.
[0013] The computing system 100 may include a central
processing unit (CPU) 21, a system memory 22 and a system
bus 23 that couples various system components including the
system memory 22 to the CPU 21. Although only one CPU is
US 2011/0078157 A1
Mar. 31, 2011
illustrated in FIG. 1, it should be understood that in some
[0016]
implementations the computing system 100 may include
hard disk 27, magnetic disk 29, optical disk 31, ROM 24 or
more than one CPU. The system bus 23 may be any of several
types of bus structures, including a memory bus or memory
RAM 25, including an operating system 35, one or more
controller, a peripheral bus, and a local bus using any of a
gram data 38, and a database system 55. The operating system
35 may be any suitable operating system that may control the
variety of bus architectures. By Way of example, and not
limitation, such architectures include Industry Standard
Architecture (ISA) bus, Micro Channel Architecture (MCA)
bus, Enhanced ISA (EISA) bus, Video Electronics Standards
Association (VESA) local bus, and Peripheral Component
Interconnect (PCI) bus also knoWn as MeZZanine bus. The
system memory 22 may include a read only memory (ROM)
24 and a random access memory (RAM) 25. A basic input/
output system (BIOS) 26, containing the basic routines that
help transfer information betWeen elements Within the com
puting system 100, such as during start-up, may be stored in
A number of program modules may be stored on the
application programs 36, an opinion search engine 60, pro
operation of a netWorked personal or server computer, such as
[email protected] XP, Mac OS® X, Unix-variants (e. g., Linux®
and BSD®), and the like. The opinion search engine 60 Will
be described in more detail With reference to FIG. 2 in the
paragraphs beloW.
[0017]
A user may enter commands and information into
the computing system 100 through input devices such as a
keyboard 40 and pointing device 42. Other input devices may
include a microphone, joystick, game pad, satellite dish,
scanner, or the like. These and other input devices may be
connected to the CPU 21 through a serial port interface 46
the ROM 24.
coupled to system bus 23, but may be connected by other
[0014] The computing system 100 may further include a
hard disk drive 27 for reading from and Writing to a hard disk,
a magnetic disk drive 28 for reading from and Writing to a
removable magnetic disk 29, and an optical disk drive 30 for
reading from and Writing to a removable optical disk 31, such
interfaces, such as a parallel port, game port or a universal
serial bus (U SB).A monitor 47 or other type of display device
may also be connected to system bus 23 via an interface, such
as a video adapter 48. In addition to the monitor 47, the
the magnetic disk drive 28, and the optical disk drive 30 may
computing system 100 may further include other peripheral
output devices such as speakers and printers.
[0018] Further, the computing system 100 may operate in a
be connected to the system bus 23 by a hard disk drive inter
face 32, a magnetic disk drive interface 33, and an optical
netWorked environment using logical connections to one or
more remote computers 49. The logical connections may be
drive interface 34, respectively. The drives and their associ
ated computer-readable media may provide nonvolatile stor
age of computer-readable instructions, data structures, pro
gram modules and other data for the computing system 100.
any connection that is commonplace in o?ices, enterprise
Wide computer netWorks, intranets, and the Internet, such as
as a CD ROM or other optical media. The hard disk drive 27,
local area netWork (LAN) 51 and a Wide area netWork (WAN)
52.
[0015] Although the computing system 100 is described
[0019]
herein as having a hard disk, a removable magnetic disk 29
computing system 100 may be connected to the local netWork
and a removable optical disk 31, it should be appreciated by
those skilled in the art that the computing system 100 may
also include other types of computer-readable media that may
be accessed by a computer. For example, such computer
readable media may include computer storage media and
communication media. Computer storage media may include
volatile and non-volatile, and removable and non-removable
media implemented in any method or technology for storage
of information, such as computer-readable instructions, data
structures, program modules or other data. Computer storage
media may further include RAM, ROM, erasable program
51 through a netWork interface or adapter 53. When used in a
mable read-only memory (EPROM), electrically erasable
programmable read-only memory (EEPROM), ?ash memory
or other solid state memory technology, CD-ROM, digital
versatile disks (DVD), or other optical storage, magnetic
cassettes, magnetic tape, magnetic disk storage or other mag
When using a LAN netWorking environment, the
WAN netWorking environment, the computing system 100
may include a modem 54, Wireless router or other means for
establishing communication over a Wide area netWork 52,
such as the Internet. The modem 54, Which may be internal or
external, may be connected to the system bus 23 via the serial
port interface 46. In a netWorked environment, program mod
ules depicted relative to the computing system 100, or por
tions thereof, may be stored in a remote memory storage
device 50. It Will be appreciated that the netWork connections
shoWn are exemplary and other means of establishing a com
munications link betWeen the computers may be used.
[0020] It should be understood that the various technolo
gies described herein may be implemented in connection With
hardWare, softWare or a combination of both. Thus, various
technologies, or certain aspects or portions thereof, may take
the form of program code (i.e., instructions) embodied in
netic storage devices, or any other medium Which can be used
to store the desired information and Which can be accessed by
tangible media, such as ?oppy diskettes, CD-ROMs, hard
the computing system 100. Communication media may
embody computer readable instructions, data structures, pro
drives, or any other machine-readable storage medium
Wherein, When the program code is loaded into and executed
gram modules or other data in a modulated data signal, such
by a machine, such as a computer, the machine becomes an
as a carrier Wave or other transport mechanism and may
apparatus for practicing the various technologies. In the case
of program code execution on programmable computers, the
computing device may include a processor, a storage medium
include any information delivery media. The term “modu
lated data signal” may mean a signal that has one or more of
its characteristics set or changed in such a manner as to
encode information in the signal. By Way of example, and not
limitation, communication media may include Wired media
such as a Wired netWork or direct-Wired connection, and
Wireless media such as acoustic, RF, infrared and other Wire
less media. Combinations of any of the above may also be
included Within the scope of computer readable media.
readable by the processor (including volatile and non-volatile
memory and/or storage elements), at least one input device,
and at least one output device. One or more programs that
may implement or utiliZe the various technologies described
herein may use an application programming interface (API),
reusable controls, and the like. Such programs may be imple
mented in a high level procedural or object oriented program
US 2011/0078157 A1
ming language to communicate With a computer system.
However, the pro gram(s) may be implemented in assembly or
machine language, if desired. In any case, the language may
be a compiled or interpreted language, and combined With
Mar. 31, 2011
specify a particular product or product category (e.g., object)
that may be of interest to the user. After receiving the speci?ed
product or product category, the opinion search engine 60
may access all of Web pages that are available on the Internet
hardWare implementations.
and determine Whether each Web page contains opinion data
[0021]
related to the speci?ed product or product category. Upon
determining that the Web page includes opinion data related
to the speci?ed product or product category, the opinion
FIG. 2 illustrates a How diagram of a method for
implementing an opinion search engine in accordance With
one or more implementations of various techniques described
herein. The folloWing description of How diagram 200 is
made With reference to computing system 100 of FIG. 1 in
accordance With one or more implementations of various
techniques described herein. It should be understood that
While the operational ?oW diagram 200 indicates a particular
order of execution of the operations, in some implementa
tions, certain portions of the operations might be executed in
a different order. In one implementation, the method for
implementing an opinion search engine may be performed by
the opinion search engine 60.
[0022] At step 210, the opinion search engine 60 may col
search engine 60 may store the corresponding Web page and
the information on the Web page in the opinion database. In
one implementation, step 210 may be executed by an opinion
data craWler module that may be part of the opinion search
engine 60. Although the opinion data has been described as
being related to products or product categories, it should be
understood that throughout this document the opinion data
may refer to any opinion about any subject matter.
[0025] At step 220, the opinion search engine 60 may
matter of interest that may be displayed on the Internet. In one
extract metadata from each Webpage stored on the opinion
database. The metadata may include information pertaining
to the opinion data on the Webpage stored on the opinion
database. In one implementation, the information pertaining
to the opinion data may include a date and time in Which the
implementation, the opinion search engine 60 may ?rst
opinion Was posted on the Webpage, a name or user identi?
access every Website and Web page available on the Internet.
cation of the user Who posted the opinion, a product pertain
ing to the opinion, a product brand pertaining to the opinion,
lect opinion data from the Internet. The opinion data may
include revieWs or ratings pertaining to an object or subject
The Web pages available on the Internet may include blogs,
neWsgroups, forums and other forms of Web-based informa
tion that may be accessible by the opinion search engine 60
via the Internet. After accessing each Web page, the opinion
search engine 60 may then determine Whether the Web page
contains opinion data about an object. Opinion data about the
objects may relate to revieWs about consumer products, vaca
tion locations or any other subject matter that may be of
interest to a user. Upon determining that the Web page
the text of the opinion, a sentiment polarity of the opinion or
any other subject matter pertaining to the opinion. The opin
ion search engine 60 may determine a product’s brand name
by using a ?nite state machine.
[0026] In order to determine a product’s brand name using
a ?nite state machine, the opinion search engine 60 may use
a Word breaker to generate or create a sequence of terms
contained in a product name as it is listed on a Web page.
includes opinion data, the opinion search engine 60 may store
Given a ?nite state machine having states and transitions
the corresponding Web page and the information on the Web
page in an opinion database. The Web pages and the informa
tion on the Web pages stored on the opinion database include
represented by terms and grammar, and an input term
a subset of the Web pages available on the Internet. In one
to ?nd a minimal set of modi?cations that may be made to the
implementation, the opinion database may be stored on a
?nite state machine such that the input term sequence is
server such that the server may be accessible via the Internet.
acceptable. In one implementation, the opinion search engine
The opinion database may be stored on the database system
60 may run the algorithm iteratively and update the ?nite state
machine such that it contains the input term sequence. The
55 as described in FIG. 1.
[0023]
In one implementation, in order to determine
Whether the Web page contains opinion data, the opinion
search engine 60 may check hoW many opinion Words are
contained in a Web page. If the number of opinion Words
exceeds a prede?ned threshold, the opinion search engine 60
may describe the Web page as having opinion data. Opinion
Words may include Words that are usually used to express
opinions (e.g., “good”, “exciting” and “bad”). In one imple
mentation, the opinion Words may be manually provided by
both human editors and machine learning algorithms. In prac
tical applications, machines may be used to generate opinion
sequence (e.g, product) that is not accepted by the ?nite state
machine, the opinion search engine 60 may use an algorithm
opinion search engine 60 may update the ?nite state machine
through induction such that the transition represented by the
terms may be changed into transitions represented by gram
mar, and neW transitions and neW states may be added to the
?nite state machine if they Were not previously represented on
the ?nite state machine. The output of algorithm is a ?nite
state machine consisting of multiple state sequences such that
each state sequence is a product name. The opinion search
engine 60 may then use the updated ?nite state machine to
predict if an input term sequence corresponds With a brand
name.
Words and then human editors may manually check or ?lter
them in order to ensure the quality of the opinion Words.
[0027] Upon extracting the metadata from the Web pages in
the opinion database, the opinion search engine 60 may deter
Although Web pages having opinion data have been described
as being identi?ed by determining hoW many opinion Words
are contained in the Web pages, it should be noted that in other
mine Whether the metadata corresponds to a particular data
property of a product ontology. The data properties of a prod
uct ontology may represent categories or de?nitions for the
implementations various algorithms may be used to deter
metadata. In one implementation, the opinion search engine
mine Whether the opinion data is contained on the Web pages.
60 may receive a list of key Words for each data property or
[0024] In another implementation, the opinion search
engine 60 may collect only targeted or speci?c opinion data
category. The opinion search engine 60 may then use a fuZZy
match algorithm to locate the key Words that may be listed in
from the Internet as opposed to all of the opinion data avail
the metadata and assign the unstructured metadata to a cor
able on the Internet. In this implementation, a user may
responding data property. The product ontology may provide
US 2011/0078157 A1
Mar. 31, 2011
an e?icient mechanism to build an association among opinion
revieWs are similar because they pertain to the same product
data and their corresponding properties. In one implementa
or object. As such, the opinion search engine 60 may combine
tion, the product ontology may be based on a universal ontol
or aggregate the metadata related to the tWo similar revieWs.
ogy model such that the raW data on any data source, such as
[0032] At step 240, the opinion search engine 60 may cat
egorize or organize the opinion data in the aggregated opinion
a Webpage, may be mapped to a corresponding data property.
In this manner, the product ontology may provide a scalable
architecture such that information from various Web sites and
data sources may be more easily integrated by the opinion
search engine 60.
[0028] In one implementation, the product ontology may
be de?ned by summarizing the details of product speci?ca
tions published by manufacturer Websites, shopping sites,
revieW sites and other related data sources. The ontology may
be de?ned in OWL, a standard ontology language published
by the World Wide Web Consortium (W3 C). OWL may facili
tate a greater machine interpretability of Web content than
that supported by Extensible Markup Language @(ML),
Resource Description Framework (RDF), RDF Schema
(RDFS), DARPA Agent Markup Language (DAML) by pro
viding additional vocabulary along With a formal semantics.
Depending on the extensibility of OWL, the product ontology
de?nition may be extended With additional ontology, prop
erty and relationship de?nition.
[0029] After identifying the keyWords related to the meta
data on each Webpage, the opinion search engine 60 may store
the metadata of the opinion data in an opinion metadata
database. In one implementation, the opinion metadata data
base may be structured in a manner such that the information
pertaining to the opinion data may be easily queried. The
opinion metadata database may be structured according to the
data properties of the product ontology. For example, the
metadata database may store information pertaining to the
opinion data in a spreadsheet such that each column of the
spreadsheet may de?ne a different property of the opinion
data. In one implementation, the text information reciting a
user’s revieW or opinion of a product may not be stored in a
structured format since the text provided by each user may
differ greatly and may not be easily queried.
[0030] At step 230, the opinion search engine 60 may
aggregate the opinion data in the opinion metadata database.
As such, the opinion search engine 60 may identify and group
all of the opinion data obtained from various Web pages on the
Internet pertaining to the same subject matter. In order to
aggregate the opinion data, the opinion search engine 60 may
determine Whether tWo or more opinion data in the opinion
metadata database are duplicative. In one implementation,
duplicative opinion data may include instances When a single
user Writes the same revieW for a single product on tWo
metadata database according to a taxonomy of hierarchal
categories. In one implementation, the opinion search engine
60 may determine the taxonomy of hierarchal categories by
evaluating the category taxonomies of the Web pages from
Which the opinion data in the opinion database may have been
obtained. In evaluating the category taxonomies of each Web
page, the opinion search engine 60 may leverage each Web
page’s link paths and site map information that may be hidden
in source sites. By leveraging each Web page’s link paths and
site map information, the opinion search engine 60 may
reduce and consolidate the taxonomy of hierarchal categories
of each Web page and resolve the inconsistencies betWeen the
different category taxonomies of different Web pages. In other
Words, the opinion search engine 60 may create a neW tax
onomy based on the consolidated taxonomy of hierarchal
categories from the different Web pages. In one implementa
tion, after determining the neW taxonomy, the opinion search
engine 60 may modify the opinion metadata database such
that it is organized according to the neW taxonomy. This
neWly organized opinion metadata database may be referred
to as a categorized opinion database. For example, on Website
“A,” there may be a “vehicle” category that contains a list of
car model manufactures like “BMW”, “Toyota,” etc. as sub
categories. On Website “B,” there may be an “automotive”
category that contains sub-categories that are similar With
those of the “vehicle” category of Website “A.” The opinion
search engine 60 may then assume that “vehicle” and “auto
motive” most likely correspond With a similar concept,
thereby resolving the inconsistencies betWeen different cat
egory taxonomies on different Web pages.
[0033] At step 250, the opinion search engine 60 may rank
or recommend products, product categories, and users’ opin
ions based on the categorized opinion data as determined in
step 240. In addition to ranking or recommending product
categories and users’ opinions, the opinion search engine 60
may display the product categories and users’ opinions
according to various aspects of the corresponding categorized
opinion data. A feW of the rank, recommendation and display
features of the opinion search engine 60 are described beloW.
Ranking Products Based on Hotness Score
[0034] In one implementation, the opinion search engine
60 may rank the products in the categorized opinion database
different Web pages. Upon determining that opinion data are
duplicative, the opinion search engine 60 may remove the
based on a hotness score. The score may be determined by
duplicative opinion data from the opinion metadata database.
[0031] In aggregating the opinion data in the opinion meta
data database, the opinion search engine 60 may determine
Whether the opinion data pertains to the similar subject matter
through log, revieW information, revieWer information, and
item selling information. The query log, click-through log,
revieW information, revieWer information, and item selling
or objects. For instance, a ?rst user may add opinion revieW to
a Web page about a certain product. Here, the ?rst user may
refer to the certain product according to its speci?c name
Which may include the manufacturer’s name and the prod
uct’s version number. A second user may also add opinion
revieW about the same product on the same or a different Web
page. The second user, hoWever, may refer to the certain
product according to its generic name Without reference to the
combining different information such as a query log, a click
information may all be available to the opinion search engine
60 on the categorized opinion database as described in step
240.
[0035]
The query log may refer to a number of query
requests by one or more users on the Internet. In one imple
mentation, the query log may indicate the interest level in one
or more particular items. The query log may be available from
an Internet search engine.
[0036] The click-through log may describe the Webpage
manufacturer’s name or the product’s version number. The
links that a user may have clicked While sur?ng the Internet.
opinion search engine 60, may determine that these tWo
For instance, after receiving Internet search results for a prod
US 2011/0078157 A1
uct of interest, the user may not click a link for each and every
product that Was returned as a search result. In contrast, the
user may only click the search results that may be related to
the products, services or opinions that may be of interest to
him. The search results clicked by the user may be recorded
on the click-through log.
[0037] The revieW information may include the number of
revieWs or opinions that may exist for each product of interest
and a trend of the number of revieWs for each product of
interest. The opinion search engine 60 may determine the
trend of the number of revieWs based the frequency of the
addition of neW revieWs over a period of time. For example, if
tWo opinions Were added in month 1 for product X and ten
opinions Were added in month 2 for product X, the opinion
search engine 60 may determine that the trend is rising
because more revieWs are being added With respect to time.
The opinion search engine 60 may also distinguish the trend
data according to positive opinion trends, negative opinion
trends and the like.
[0038] The revieWer information may describe the user
Who adds opinions on the Internet. The revieWer information
may include the number of revieWers commenting or adding
an opinion for a product, the demographic information per
taining to the user (e. g., gender, age and location coverage of
user) and the like.
Mar. 31, 2011
opinion search engine 60 estimate the authority of each user
based on Whether the user is an editor on the Website, a
product specialist or the like.
[0042] In another implementation, various Websites may
provide a rating for each user. In this case, the opinion search
engine 60 may estimate the authority of each user based on
the rating provided by the Website. If the opinion search
engine 60 determines that an opinion has been added from a
user With a higher authority, the opinion search engine 60 may
give that user’s opinion more Weight in determining the uni
?ed opinion score. In addition to estimating the authority of
each user, the opinion search engine 60 may also estimate the
authority of each Website containing user opinions.
[0043] The opinion search engine 60 may also evaluate the
revieW provided by a user in order to calculate an opinion
score. Here, the opinion search engine 60 may analyZe the
revieW provided by the user and determine Whether the revieW
includes characteristics that indicate that it is a positive
revieW about the product or a negative revieW about the prod
uct.
Rank Products Based on a Recommendation Score
[0044]
The opinion search engine 60 may also rank the
products according to a recommendation score. The opinion
search engine 60 may calculate the recommendation score
based on the volume of the opinions or revieWs provided for
[0039] The item selling information may describe the num
ber of products that have been sold to or purchased by the
users. The opinion search engine 60 may then determine a
a particular product and the percentage of positive opinions or
revieWs related to the particular product. In this manner, the
hotness score for the product revieWed by the users based on
the query log, click-through log, revieW information,
its recommendation score. The recommendation score may
be based on Whether the total number of users Who added a
revieWer information and item selling information. Based on
the hotness score for the various products revieWed by the
positive opinion on the Internet about the particular product is
users, the opinion search engine 60 may also rank product
categories according to a hotness score that may be computed
based on the hotness scores of all of the products in each
product category. In one implementation, the opinion search
engine 60 may compute the hotness scores for each product in
real-time While the user is online. In other implementations,
the opinion search engine 60 may compute the hotness scores
for each product While the user is of?ine using the informa
opinion search engine 60 may rank each product according to
greater than the number of users Who added a negative opin
ion about the particular product. For example, suppose there
are 100 revieWs for product X and 10 revieWs for productY. 49
percent of the revieWs for product X are positive, and 50
percent of the revieWs for productY are positive. The opinion
search engine 60 may determine the recommendation score
based on the number of users Who Wrote a positive revieW as
opposed to the percentage of the users Who had a positive
revieW for a product. In the preceding example, product X
tion previously obtained from the users on the Internet.
may have a higher recommendation score as compared to
Rank Products by User Opinion
productY because product X had a total of 49 positive revieWs
While product Y only had a total of 5 positive revieWs.
[0040]
In yet another implementation, the opinion search
engine 60 may rank the products according to the users’
opinions or an opinion score about the product. In order to
rank the products according to the users’ opinions about the
product, the opinion search engine 60 may evaluate the rat
ings and the ranking of various products as provided on
various Websites. For instance, on some shopping Websites,
products may be ranked based on average rating scores pro
vided by Internet users. Since each shopping Website may
employ its oWn rating system, the opinion search engine 60
may combine the ratings for a product from each Website to
determine a uni?ed opinion score for each product.
[0041]
Displaying Products by Dynamic Groups
[0045] The opinion search engine 60 may display products
according to dynamic groups. Here, the opinion search
engine 60 may ?lter the products according to the dynamic
groups. The opinion search engine 60 may ?lter products
according to a product brand, a manufacturer of the product,
or features of the product. In one implementation, the
dynamic groups may be de?ned by a user by specifying to the
opinion search engine 60 a percentage of positive revieWs for
a feature of the product. For example, a product may include
features such as a user manual, noise, resolution and Weight.
The opinion search engine 60 may calculate the
The opinion search engine 60 may display a WindoW to the
uni?ed opinion score based on a total number of users Writing
a revieW about a product, the rating score of each user if a
user such that the user may specify hoW to ?lter the products
in the categorized opinion database. The user may specify a
rating score is available, Whether the revieW is positive or
negative if a rating score is unavailable, an authority estimate
for each user, an authority estimate for each Website contain
ing user revieWs and the like. In one implementation, the
percentage of positive revieWs for each feature of the product.
In one implementation, the user may be interested in a prod
uct having high ratings for its user manual. As such, the user
may specify to the opinion search engine 60 that the percent
US 2011/0078157 A1
Mar. 31, 2011
age of positive reviews pertaining to the product’s user
manual Will be 70% or greater. Accordingly, the opinion
search engine 60 may remove all of the products having less
than 70% positive revieWs for its manual. In one implemen
tation, the opinion search engine 60 may determine Whether
the revieWs for each feature of a product include positive or
pertaining to the positive and negative opinions is not distin
guishable. HoWever, if there is a large difference betWeen the
percentage of positive opinions and the percentage of nega
tive opinions pertaining to the product feature, the opinion
negative revieWs based on a calculated opinion score as
[0051] The opinion search engine 60 may also use other
information pertaining to the product, such as speci?cation
information about the product, to determine Whether the
described above.
search engine 60 may determine that the product feature is a
distinguishing feature.
Recommending Related Products
product includes one or more distinguishing features. In one
[0046]
implementation, the product speci?cation for one product in
The opinion search engine 60 may also identify
products that are related to each other. The related products
may have similar properties, features and/or components. For
example, different products may use a similar memory device
in order to store information (i.e., SD memory card). In this
example, the different products using similar memory
devices may be associated With each other as related prod
ucts. As such, the opinion search engine 60 may associate a
product With one or more other products having the same or
similar values of related properties in order to recommend
a product category may indicate that the product costs sig
ni?cantly more than the other products in the same category.
Here, the price of the product may be described as a distin
guishing feature because of the disparity betWeen the price of
the product as compared to prices of similar products. Simi
larly, if other features provided in a product’s speci?cation
differ signi?cantly from the features of other products that fall
Within the same category of the product, the opinion search
engine 60 may determine that these differing features may be
products.
distinguishing features.
Recommending Competitor Products
Query Products Based on Positive/Negative Percentage in
[0047] The opinion search engine 60 may also recommend
competitor products to a user. Here, the opinion search engine
Features
[0052]
The opinion search engine 60 may also receive a
60 may analyZe a user’s click-through log to identify one or
request from a user to query products based on the positive
more Web pages related to a speci?c product that a user may
and negative opinion percentages in features as described
above. In this implementation, the opinion search engine 60
may evaluate the percentage of positive opinions and the
have accessed by clicking on a search result. Upon analyZing
the click-through log, the opinion search engine 60 may iden
tify one or more competitor products of the product Web
pages accessed by the user. The opinion search engine 60 may
then recommend these competitor products to the user.
[0048] In order to determine Which products may be com
petitors to the speci?c product, the opinion search engine 60
may analyZe the click-through log and determine Which prod
percentage of negative opinions on each feature to determine
Whether the feature is a distinguishing feature. The opinion
search engine 60 may then list the products according to the
products having the highest discrepancy betWeen the percent
age of positive opinions and the percentage of negative opin
ions on the feature.
uct Web pages Were accessed by multiple users as indicated in
the multiple users’ click-through logs. In one implementa
tion, the opinion search engine 60 may determine that prod
Compare Products Based on Opinions
ucts may be competitors if multiple users access different
[0053]
product Websites after performing a query for a speci?c prod
comparison betWeen tWo or more products. The comparison
The opinion search engine 60 may also display a
uct or for a generic product category.
betWeen the products may compare the distinguishing fea
tures betWeen the tWo or more products. The opinion search
Identifying Distinguishing Features of a Product
engine 60 may identify the distinguishing features according
[0049]
to the method for identifying distinguishing features of a
product as described above.
The opinion search engine 60 may also identify one
or more distinguishing features of a product based on the
categoriZed opinion database. The opinion search engine 60
may then display these distinguishing features to a user. In
one implementation, the opinion search engine 60 may iden
tify the distinguishing features of a product by analyZing the
number of users commenting on the feature, the positive and
negative opinion percentages of the user comments on the
feature and a differentiation in feature level betWeen each
product falling in the same category in the categorized opin
ion database.
[0050]
In one implementation, if the number of users com
menting on a speci?c feature is large, the opinion search
engine 60 may determine that the speci?c feature is a distin
guishing feature. The opinion search engine 60 may also use
the positive and negative opinion percentages of the user
comments about the feature to determine Whether a feature is
Display and Compare Opinion Trend
[0054] The opinion search engine 60 may also organiZe
opinion data according its trend. Here, the opinion search
engine 60 may evaluate the trend of the opinion data or
revieWs for a particular product or a product feature over a
period of time. In one implementation, if the number of opin
ions pertaining to the product or the product feature is increas
ing, the opinion search engine 60 may display an arroW point
ing up next to the corresponding product or product feature to
indicate that the trend for the corresponding opinion data
about the product or product feature is rising. If, hoWever, the
number of opinions pertaining to the product or the product
feature is decreasing, the opinion search engine 60 may dis
play an arroW pointing doWn next to the corresponding prod
distinguishable. Here, if the percentage of positive opinions
and the percentage of negative opinions are relatively similar,
uct or product feature to indicate that the trend for the corre
the opinion search engine 60 may determine that the feature
falling.
sponding opinion data about the product or product feature is
US 2011/0078157 A1
[0055]
The opinion search engine 60 may also display the
trend of the opinion data for a single product over a period of
time on a graph as shoWn in FIG. 3. FIG. 3 illustrates the trend
of the opinion data over a time period of one year. As such, the
total number of opinions for each month during one year is
illustrated on the graph 300. The trend of the total number of
opinions represented by the line 3 1 0. In addition to displaying
the trend of the opinion data for a single product, the opinion
search engine 60 may compare the trend of the opinion data
over a period of time for multiple products. In this manner, the
opinion data trend for each of the multiple products may be
represented by different lines on a graph similar to that as
illustrated in FIG. 3.
VieW Opinion Change Data
[0056]
The opinion search engine 60 may determine the
volume of change in the number of user comments for a
product, the number of positive comments for a product, the
number of negative comments for a product, the percentage of
user comments for a product, the percentage of positive com
Mar. 31, 2011
[0059] Although the subject matter has been described in
language speci?c to structural features and/or methodologi
cal acts, it is to be understood that the subject matter de?ned
in the appended claims is not necessarily limited to the spe
ci?c features or acts described above. Rather, the speci?c
features and acts described above are disclosed as example
forms of implementing the claims.
What is claimed is:
1. A computer-readable storage medium having stored
thereon computer-executable instructions Which, When
executed by a computer, cause the computer to:
collect opinion data about one or more objects from the
Internet;
extract metadata about the opinion data from the opinion
data;
remove duplicate metadata from the metadata to generate a
resulting metadata;
categoriZe the resulting metadata for similar objects
according to one or more taxonomies from one or more
Websites on the Internet; and
rank the similar objects based on the categoriZed metadata.
2. The computer-readable storage medium of claim 1,
ments for a product, the percentage of negative comments for
a product and the like. The opinion search engine 60 may
determine the change With respect to a time period (e.g.,
Week, month) as de?ned by a user. After determining the
change in the various types of comments for a product, the
Wherein the computer-executable instructions Which, When
opinion search engine 60 may display the opinion change
Which, When executed by a computer, cause the computer to:
data to the user. FIG. 4 illustrates the opinion data change in
volume and percentage from Month 1 to Month 2. The opin
ion search engine 60 may also display the opinion data change
information for multiple products on the same graph. In addi
tion to determining the opinion change data for a product, the
opinion search engine 60 may also determine the opinion
change data for features of a product.
executed by the computer, cause the computer to collect the
opinion data comprises computer-executable instructions
access one or more Webpages on the Internet;
determine Whether the Webpages contain the opinion data;
and
store the Webpages containing the opinion data in a data
base.
3. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to collect the
Rank Opinion Data
opinion data comprises computer-executable instructions
[0057]
Which, When executed by a computer, cause the computer to:
receive a description of the objects;
The opinion search engine 60 may also rank the
opinion data that a user has added on the Internet. In one
implementation, the opinion search engine 60 may rank the
opinion data based on an overall ranking algorithm. The
overall ranking algorithm may consider the quality of the
opinion data, the authority of the user expressing the opinion,
the authority of the Website from Which the opinion Was
obtained and the like in ranking the opinion data. The opinion
search engine 60 may also rank opinion data based on feature
ranking algorithm. The feature ranking algorithm may rank
the opinion data based on the same information considered by
the overall ranking algorithm along With information pertain
ing to hoW much of the opinion data is about a feature of
interest. The feature of interest may be de?ned by the user
performing a search of opinion data on the opinion search
engine 60.
Generate Demographic Statistics of Opinion Data
[0058] The opinion search engine 60 may also generate
demographic statistics of the opinion data. The demographic
statistics may include location, gender, age etc. of the user
Who added the opinion data. The opinion search engine 60
may provide the demographic details of the opinion data to its
users. The opinion search engine 60 may also provide opinion
data search results based on the similarities betWeen demo
graphic details of the opinion data and the demographic
details of the user.
access one or more Webpages on the Internet;
determine Whether the Webpages contain the opinion data
about the objects; and
store the Webpages containing the opinion data about the
objects in a database.
4. The computer-readable storage medium of claim 1,
Wherein the objects comprise one or more products, one or
more product categories, vacation locations, revieWs or any
subject matter of interest.
5. The computer-readable storage medium of claim 1,
Wherein the metadata comprises a date and a time of the
opinion data, an author of the opinion data, a rating of the
objects, a sentiment polarity of the objects, a revieW of the
objects or combinations thereof.
6. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to extract the
metadata comprises computer-executable instructions Which,
When executed by a computer, cause the computer to:
determine Whether the metadata corresponds to one or
more categories; and
store the metadata according to the categories.
7. The computer-readable storage medium of claim 6,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to determine
Whether the metadata corresponds to the categories com
US 2011/0078157 A1
Mar. 31, 2011
prises computer-executable instructions Which, When
identify one or more distinguishing features of the similar
executed by a computer, cause the computer to:
receive one or more key Words for each category;
rank the similar objects based on the distinguishing fea
locate the key Words in the metadata using a fuzzy match
algorithm; and
assign the metadata having the key Words to a correspond
ing category.
8. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to categorize
metadata comprises computer-executable instructions Which,
When executed by a computer, cause the computer to:
consolidate the taxonomies; and
organize the metadata according to the consolidated tax
onomies.
9. The computer-readable storage medium of claim 8,
Wherein the taxonomies are consolidated by leveraging each
Website’s one or more link paths and site map information.
10. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to rank the
similar objects comprises computer-executable instructions
Which, When executed by a computer, cause the computer to:
calculate a score for each categorized metadata using a
query log, a click-through log, revieW information,
revieWer information, and purchase history related to the
similar objects; and
rank the similar objects according to the score.
11. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to rank the
similar objects comprises computer-executable instructions
Which, When executed by a computer, cause the computer to:
combine one or more ratings for the similar objects that are
listed in the categorized metadata; and
rank the similar objects according to the ratings.
12. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to rank the
similar objects comprises computer-executable instructions
Which, When executed by a computer, cause the computer to:
determine Whether one or more revieWs about the similar
objects are positive or negative;
estimate one or more authority values for one or more
authors of the revieWs and for one or more Webpages
Where the revieWs are located; and
rank the similar objects based on the revieWs, the authority
values, or combinations thereof.
13. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
objects; and
tures.
15. The computer-readable storage medium of claim 14,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to identify the
distinguishing features comprises computer-executable
instructions Which, When executed by a computer, cause the
computer to:
determine that a feature of the similar objects is distin
guishing When a number of revieWs pertaining to the
feature exceeds a number of revieWs pertaining to other
features of the similar objects by a ?rst predetermined
value;
determine that the feature of the similar objects is distin
guishing When a percentage of positive revieWs about
the feature exceeds a percentage of negative revieWs
about the feature by a second predetermined value;
determine that the feature of the similar objects is distin
guishing When a discrepancy betWeen a ?rst value of the
feature for a ?rst object and a second value of the feature
for a second object of the similar objects exceeds a third
predetermined value; or
combinations thereof.
16. A computer system, comprising:
at least one processor; and
a memory comprising program instructions executable by
the at least one processor to:
collect opinion data about one or more objects from the
Internet;
extract metadata about the opinion data from the opinion
data;
remove duplicate metadata from the metadata to gener
ate a resulting metadata;
categorize the resulting metadata for similar objects
according to one or more taxonomies from one or
more Websites on the lntemet;
receive one or more features about the similar objects
and a minimum percentage of positive revieWs for
each feature; and
display the similar objects having the features and meet
ing the minimum percentage of positive revieWs.
17. The computer system of claim 16, Wherein the memory
further comprises program instructions executable by the at
least one processor to display a trend indicating Whether a
number of revieWs about one of the similar objects is rising or
falling over a time period.
18. The computer system of claim 16, Wherein the memory
executed by the computer, cause the computer to rank the
further comprises program instructions executable by the at
similar objects comprises computer-executable instructions
least one processor to display a change in a number revieWs
for the similar objects over a time period.
Which, When executed by a computer, cause the computer to:
calculate one or more recommendation scores based on a
19. The computer system of claim 16, Wherein the memory
further comprises program instructions executable by the at
volume of one or more revieWs about the similar objects
and one or more percentages of positive revieWs about
least one processor to recommend one or more related objects
the similar objects; and
having similar properties, features, components or combina
rank the similar objects based on the recommendation
scores.
14. The computer-readable storage medium of claim 1,
Wherein the computer-executable instructions Which, When
executed by the computer, cause the computer to rank the
similar objects comprises computer-executable instructions
Which, When executed by a computer, cause the computer to:
tions thereof, as the similar objects.
20. A method for creating an opinion search engine, com
prising:
collecting opinion data about one or more objects from the
lntemet;
extracting metadata about the opinion data from the opin
ion data;
US 2011/0078157 A1
removing duplicate metadata from the metadata to generate a resulting metadata;
categorizing resulting metadata for similar objects accord
ing to one or more taxonomies from one or more Web
sites on the lntemet;
receiving one of the objects; and
Mar. 31, 2011
ranking, using amicroprocessor, the opinion data about the
one of the objects based on a quality of one or more
reVieWs, one or more authority Values of one or more
authors of the reVieWs and one or more Webpages from
Which the reVieWs exist.
*
*
*
*
*
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