HPE IDOL Collaborative Classifier 11.1 User Guide

HPE IDOL Collaborative Classifier 11.1 User Guide
HPE IDOL Collaborative Classifier
Software Version: 11.1
User Guide
Document Release Date: February 2017
Software Release Date: October 2016
User Guide
Legal Notices
Warranty
The only warranties for Hewlett Packard Enterprise Development LP products and services are set forth in
the express warranty statements accompanying such products and services. Nothing herein should be
construed as constituting an additional warranty. HPE shall not be liable for technical or editorial errors or
omissions contained herein.
The information contained herein is subject to change without notice.
Restricted Rights Legend
Confidential computer software. Valid license from HPE required for possession, use or copying. Consistent
with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation, and
Technical Data for Commercial Items are licensed to the U.S. Government under vendor's standard
commercial license.
Copyright Notice
© Copyright 2017 Hewlett Packard Enterprise Development LP
Trademark Notices
Adobe™ is a trademark of Adobe Systems Incorporated.
Microsoft® and Windows® are U.S. registered trademarks of Microsoft Corporation.
UNIX® is a registered trademark of The Open Group.
This product includes an interface of the 'zlib' general purpose compression library, which is Copyright ©
1995-2002 Jean-loup Gailly and Mark Adler.
Documentation updates
The title page of this document contains the following identifying information:
l
l
l
Software Version number, which indicates the software version.
Document Release Date, which changes each time the document is updated.
Software Release Date, which indicates the release date of this version of the software.
To check for recent software updates, go to https://softwaresupport.hpe.com/patches.
To verify that you are using the most recent edition of a document, go to
https://softwaresupport.hpe.com/manuals.
This site requires that you register for an HPE Passport and sign in. To register for an HPE Passport ID, go to
https://hpp12.passport.hpe.com/hppcf/login.do.
You will also receive updated or new editions if you subscribe to the appropriate product support service.
Contact your HPE sales representative for details.
Support
Visit the HPE Software Support Online web site at https://softwaresupport.hpe.com.
This web site provides contact information and details about the products, services, and support that HPE
Software offers.
HPE IDOL Collaborative Classifier (11.1)
Page 2 of 155
User Guide
HPE Software online support provides customer self-solve capabilities. It provides a fast and efficient way to
access interactive technical support tools needed to manage your business. As a valued support customer,
you can benefit by using the support web site to:
l
l
l
l
l
l
l
l
l
Search for knowledge documents of interest
Submit and track support cases and enhancement requests
Download software patches
Access product documentation
Manage support contracts
Look up HPE support contacts
Review information about available services
Enter into discussions with other software customers
Research and register for software training
Most of the support areas require that you register as an HPE Passport user and sign in. Many also require a
support contract.
To register for an HPE Passport ID, go to https://hpp12.passport.hpe.com/hppcf/login.do.
To find more information about access levels, go to
https://softwaresupport.hpe.com/web/softwaresupport/access-levels.
HPE IDOL Collaborative Classifier (11.1)
Page 3 of 155
User Guide
Contents
Part 1: Getting Started
11
Chapter 1: Introduction
HPE IDOL Collaborative Classifier
The HPE IDOL Platform
HPE IDOL Collaborative Classifier Components
13
13
13
13
Chapter 2: Installation and Configuration
System Requirements
Hardware Requirements
Software Requirements
Install the HPE IDOL Collaborative Classifier
Install or Reinstall HPE IDOL Collaborative Classifier
Install by using a Tomcat Application Server
Configure the Heap Size
Bootstrap the HPE IDOL Collaborative Classifier
Configure the Taxonomy Module
Enable Automatic Query Guidance (AQG)
Configure Document Viewing
Enable Secure Connections
Enable Encrypted HPE IDOL Connections
Enable Paging Managers
Enable Binary Categories
Install and Access the HPE IDOL Collaborative Classifier Client
15
15
15
15
15
15
15
16
17
18
18
18
19
19
20
20
21
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Icons
Taxonomy Module Icons
Categorization and QA Icons
Manage Binary Category Icons
Windows
Summary Window
Interface Panes
Binary Category Features
Menus
File Menu
Window Menu
Help Menu
View Menu
Taxonomy Filters Menu
Resources Menu
Module Navigation Shortcut Menu
Taxonomy Filtered Draft Shortcut Menu
23
23
24
24
25
26
26
26
26
27
27
27
27
28
28
28
28
29
HPE IDOL Collaborative Classifier (11.1)
Page 4 of 155
User Guide
Published Shortcut Taxonomy Menu
Query Tree Shortcut Taxonomy Menu
Name Restrictions
Taxonomy Module Resource Name Restrictions
Application Module Resource Name Restrictions
Chapter 4: HPE IDOL Collaborative Classifier Administration
About Resource Administration
Maintain Modules
Create a Resource
Delete a Resource
Manage Users
User Permissions
Module User Roles
Add New Users
Manage User Permissions
Delete a User
Change Default Taxonomy Properties
Edit an Application URL
Part 2: Taxonomies
30
30
30
30
31
33
33
33
33
34
34
34
35
35
36
36
36
37
39
Chapter 5: About Taxonomies
Classify Information
Create, Populate, and Use Taxonomies
Build the Taxonomy
Define Categories
Populate the Taxonomy With Documents
Use the Taxonomy
About Category Rules
About Categories and Taxonomies
Categories and Properties
Reference Categories
Resource Taxonomies
How the Taxonomy Module Handles Documents
About Databases
About Binary Categories
41
41
42
42
44
44
45
45
46
46
46
47
47
47
47
Chapter 6: Create Taxonomies
Work with Taxonomies
Open and Close Taxonomies
Opening and Closing Resource Taxonomies
Select a Taxonomy View
Edit Categories
Request Publication of a Category
Add Taxonomies
Create an Empty Taxonomy
Import a Taxonomy
49
49
49
50
50
50
50
51
51
52
HPE IDOL Collaborative Classifier (11.1)
Page 5 of 155
User Guide
Automatically Create a Taxonomy from a Database
Optional Settings for Automatic Taxonomy Creation
Prefix to Ignore
Ignore Suffix
Assign Documents
Generate Rule
53
55
55
55
55
56
Chapter 7: Author Taxonomies
Author Role
Edit a Taxonomy
Add a Category
Name Categories
Delete a Category or Subtree
Restore a Category or Subtree
Rename a Category
Search a Category Node
Move a Category Subtree
Copy a Category Subtree
Reference Category Rules
Copy a Resource Category
Edit Category Information
Add a Category Description
Change the Category Language
Add Reviewer Notes
View Reviewer Notes History
Create User-Defined Properties for a Category Node
Add a User-Defined Property
Edit a Property
Delete a User-Defined Property
Edit HPE IDOL Standard Properties
Edit Number of Results Returned by a Query
Edit Databases
Add and Maintain Users
Add a User
Edit User Privileges
Remove a User
About Filters
Create Custom Filters
Remove Custom Filters
View Published Categories
57
57
57
58
58
58
59
59
59
60
60
61
61
62
62
62
63
63
64
64
64
64
65
65
65
66
66
66
67
67
67
68
68
Chapter 8: Author Binary Categories
About Binary Taxonomy Authoring
Edit Binary Categories
Add a Binary Category
Delete a Binary Category from the Taxonomy
Edit Binary Category Information
Edit HPE IDOL Standard Properties
69
69
69
69
70
70
71
HPE IDOL Collaborative Classifier (11.1)
Page 6 of 155
User Guide
Manage Users
Chapter 9: Populate Categories
Introduction
Select a Categorization Method
Training
Query
Manual
Search a Database
Search by Saved Searches
Create a Subtaxonomy
Create a Subtaxonomy from Selected Documents
Create a Subtaxonomy from Search Results
Define Category Rules by Using Training Documents
Mark Documents for Training
Remove Documents from the Training Documents List
Train the Category
Refine a Category Rule by Using Terms and Weights
Add or Edit a Term
Keep a Term
Remove Terms
User-Defined Term Properties
About Document Best Terms
Define Category Rules by Using a Query
Build a Query Text
Enter Query Text Directly
Use the Query Text Wizard
Build a Query by Using a Field Restriction Query Type
Build a Query by using a Fuzzy Query Type
Build a Query by Using a Soundex Keyword Query Type
Build a Query by Using an Existing Rule Query Type
Append a Query
Build a Query Filter
Test a Query
About Query Text and Filter History
Handle Unsaved Data
View and Retrieve Queries and Filters
Rules for Child Categories
Manually Assign Documents
Assign Documents to Accepted
Assign Documents for Review
Assign Documents to Excluded
About the Manual Action Results Pane
About Automatic Query Guidance
Run AQG From the Search Results
Run AQG From Selected Documents
About AQG Results
HPE IDOL Collaborative Classifier (11.1)
71
73
73
73
74
75
77
78
79
79
80
80
81
81
82
82
82
83
83
84
84
85
85
85
85
86
86
88
89
90
91
93
95
95
95
96
97
97
97
97
98
98
98
99
99
99
Page 7 of 155
User Guide
Chapter 10: Populate Binary Categories
Introduction
Binary Category Training Workflow
Train Binary Categories
Define Training Criteria
Select Training Documents
Enter Training Text
Select Training Directories
Train the Category
View Terms and Weights
Binary Queries
101
101
101
101
102
102
102
103
103
103
103
Chapter 11: Manage Documents
About Document Management
Configure the Document List
Add or Remove Columns in the Document List
Change the Document List Output
Change the Documents Shown
View Documents Added Manually or Automatically
View Documents in Different Views
Paging Managers
About the Document Distribution Graph
Read the Graph
Change the Threshold
Manipulate Documents in the List
Move Documents into Another View
Remove Approved Documents from a Category
Change a Document Ranking
Recommend Documents to Other Categories
Revert a Manual Action
Mark Documents for Training
Filter the Category
105
105
106
107
108
108
108
108
109
109
110
111
111
112
112
112
113
113
113
114
Chapter 12: Review Categories
Reviewer Tasks
Verify a Category Rule
Search for Documents
View Documents and Marking for Benchmark
Remove Documents from Benchmark
Evaluate the Category Rule
Add Review Comments to a Category
Generate Category Reports
115
115
115
116
116
117
117
118
118
Chapter 13: Publish and Export a Taxonomy
About Publishing
User Roles
Publish Categories
Publish a Category
121
121
121
121
121
HPE IDOL Collaborative Classifier (11.1)
Page 8 of 155
User Guide
Publish a Root Category
Publish to a Remote Server
Publish Binary Categories
Publish a Binary Category
Publish to a Remote Server
Export a Taxonomy
Part 3: Applications
122
122
123
123
124
124
127
Chapter 14: Create Applications
About Applications
Activate the Applications Module
Work with Applications
Open and Close Applications
Refresh an Application
Add Applications
Create a New Configuration
Register an Application
129
129
130
130
130
130
131
131
132
Chapter 15: Modify Applications
Add Contexts
Add Components
Add a New Component
Share a Component
Delete Contexts or Components
Edit Contexts or Components
Edit Fields
View the XML
Preview the Application
Publish an Application
135
135
136
136
137
138
138
138
139
139
139
Part 4: Appendixes
141
Appendix A: Workflow Example
143
Appendix B: HPE Category XML Format
Introduction
XML Format
<autn:categories> (required)
<autn:category>
<autn:name> (required)
<autn:id>
<autn:parent>
<autn:refersto>
<autn:trainingelement>
<autn:simplecat>
<autn:relevancecat>
<autn:details>
145
145
145
145
146
146
146
147
147
147
148
149
149
HPE IDOL Collaborative Classifier (11.1)
Page 9 of 155
User Guide
Example Category XML Files
Send documentation feedback
HPE IDOL Collaborative Classifier (11.1)
153
155
Page 10 of 155
User Guide
Part 1: Getting Started
This section describes how to install and configure HPE ICC, and how to administer resources.
l
Introduction
l
Installation and Configuration
l
Overview of HPE IDOL Collaborative Classifier
l
HPE IDOL Collaborative Classifier Administration
HPE IDOL Collaborative Classifier (11.1)
Page 11 of 155
User Guide
HPE IDOL Collaborative Classifier (11.1)
Page 12 of 155
User Guide
Chapter 1: Introduction
Chapter 1: Introduction
This section introduces the HPE ICC and its components.
•
•
•
HPE IDOL Collaborative Classifier
13
The HPE IDOL Platform
13
HPE IDOL Collaborative Classifier Components
13
HPE IDOL Collaborative Classifier
HPE IDOL Collaborative Classifier (HPE ICC) manipulates information classification structures and
indexes. It also helps the development of search and classification applications.
For an example HPE ICC workflow, see Workflow Example, on page 143.
The HPE IDOL Platform
HPE ICC integrates with the HPE Intelligent Data Operating Layer (HPE IDOL).
HPE IDOL gathers and processes unstructured, semistructured, and structured information in any
format from multiple repositories, using HPE IDOL connectors and a global relational index. It can
automatically form a contextual understanding of the information in real time, linking disparate data
sources based on the concepts contained within them. For example, HPE IDOL can automatically link
concepts contained in an email message to a recorded phone conversation, which can be associated
with a stock trade. This information is then imported into a format that is easily searchable, adding
advanced retrieval, collaboration, and personalization to an application that integrates the technology.
For more information on HPE IDOL, see the HPE IDOL Getting Started Guide.
HPE IDOL Collaborative Classifier Components
HPE ICC consists of the following modules:
l
l
Taxonomy module. A module for creating, viewing, editing, and testing taxonomies. You can use
the Taxonomy module to classify information into categories for more efficient browsing.
Application module. A module for creating and modifying applications. With this module, you can
manage applications that work with HPE IDOL. You can structure datasheets, catalogs, and other
content to create specific navigation paths for the users.
NOTE:
The application module is disabled by default. To enable the application module, see Activate
the Applications Module, on page 130.
HPE IDOL Collaborative Classifier (11.1)
Page 13 of 155
User Guide
Chapter 1: Introduction
HPE IDOL Collaborative Classifier (11.1)
Page 14 of 155
User Guide
Chapter 2: Installation and Configuration
Chapter 2: Installation and Configuration
This section describes how to install and configure the HPE IDOL Collaborative Classifier (HPE ICC)
Taxonomy and Application modules.
•
•
•
•
System Requirements
15
Install the HPE IDOL Collaborative Classifier
15
Configure the Taxonomy Module
18
Install and Access the HPE IDOL Collaborative Classifier Client
21
System Requirements
If you are installing HPE IDOL Collaborative Classifier with HPE IDOL, you must meet the system
requirements for the HPE IDOL installation. See the HPE IDOL Server Getting Started Guide for more
information about installing HPE IDOL.
Hardware Requirements
l
HPE recommends a minimum of 1 GB of RAM for both the server and client.
Software Requirements
l
Tomcat 8 and later
l
HPE IDOL Server
Install the HPE IDOL Collaborative Classifier
See Install and Access the HPE IDOL Collaborative Classifier Client, on page 21 for the client
installation steps.
Install or Reinstall HPE IDOL Collaborative Classifier
The following installation instructions apply to both new installations and reinstallations.
Install by using a Tomcat Application Server
To install standalone HPE ICC by usingTomcat
1. If you have previously installed HPE ICC, close all HPE ICC windows and stop the Applications
service. Close your web browser.
2. (Optional) Create a backup of your iccdb/ directory if you want to reuse it.
HPE IDOL Collaborative Classifier (11.1)
Page 15 of 155
User Guide
Chapter 2: Installation and Configuration
3. Go toTomcatX/webapps/, where X is the Tomcat version number. Delete acc.war and the icc/
directory.
4. Copy the new acc.war file into TomcatX/webapps/.
5. Start the Tomcat applications server.
6. Follow the instructions to bootstrap HPE ICC. See Bootstrap the HPE IDOL Collaborative
Classifier, on the next page.
7. (Optional) If you are using a saved iccdb/ directory, stop the Tomcat applications server and copy
the saved iccdb/ directory to TomcatX/webapps/icc/, where X is the Tomcat version number.
Start the Tomcat applications server and start HPE ICC.
Configure the Heap Size
The heap size affects the memory available to the HPE IDOL Collaborative Classifier modules. A
small heap size might negatively impact performance. HPE recommends a maximum heap size of at
least 1 GB for any production use of HPE ICC.
You can configure the heap size either through the Tomcat server or the HPE ICC configuration file.
To configure the heap size through the Tomcat server
l
On Windows, through the Properties dialog box:
1. In a command-line window, navigate to the <installDir>/TomcatX/bin directory, where
installDir is the location of your Tomcat installation, and X is the Tomcat version number.
2. Run the following command:
tomcat8w.exe //ES//
The Tomcat Server Properties dialog box opens.
TIP:
You can also open the Tomcat Server Properties dialog by clicking
taskbar.
in your Windows
3. On the Java tab, change the Maximum memory pool setting to the desired heap size.
4. Click Apply, and then restart Tomcat.
l
On Windows, through the BAT file:
1. Navigate to the <installDir>/TomcatX/bin directory, where installDir is the location of
your Tomcat installation, and X is the Tomcat version number.
2. Open the catalina.bat file.
3. In JAVA_OPTS, include the -Xmx parameter and specify the desired heap size. For example:
-Xmx1024m
sets the maximum heap size to 1 GB.
4. Save the file, and then restart Tomcat.
HPE IDOL Collaborative Classifier (11.1)
Page 16 of 155
User Guide
Chapter 2: Installation and Configuration
Bootstrap the HPE IDOL Collaborative Classifier
When you bootstrap the HPE IDOL Collaborative Classifier (HPE ICC), you provide the basic system
configuration settings so that HPE ICC can install fully.
See Install and Access the HPE IDOL Collaborative Classifier Client, on page 21 for the client
installation steps.
To bootstrap the HPE IDOL Collaborative Classifier
NOTE:
If you configure the startup page for login, you are the user with administrator user privileges in
HPE ICC.
Only one administrator user is allowed.
1. In your browser, go to:
http://<hostname>:<port>/icc
where:
hostname The host on which you installed the HPE ICC modules.
port
The port number of the Tomcat server on the host.
The HPE ICC bootstrap page opens.
2. Select either the Unencrypted or Encrypted HPE IDOL Connection Type. For more information
about the Encrypted HPE IDOL Connection Type, see Enable Encrypted HPE IDOL
Connections, on page 19.
3. Select either the HTTP (the default) or HTTPS protocol. HTTPS enables secure connections
between HPE ICC and HPE IDOL Server. See Enable Secure Connections, on page 19.
4. Enter the hosts and ports of the HPE Community Server (used for authentication), HPE Search
Server, HPE Category Server, and HPE View Server. You can either enter the HPE IDOL Server
host and port for all these values, or identify the components directly.
If you do not know the host name, use the host name or IP address of the machine on which HPE
ICC is running. If you do not know the correct port numbers, contact your system administrator.
NOTE:
By default, HPE ICC publishes taxonomies to the HPE Category Server. You can also
publish taxonomies to remote servers. See Taxonomies, on page 39.
5. In the Admin User and Password fields, enter your user name and password .
If you do not have a user name and password, use the default user, abcadmin, for both fields.
6. Click Bootstrap.
The launch screen opens.
HPE IDOL Collaborative Classifier (11.1)
Page 17 of 155
User Guide
Chapter 2: Installation and Configuration
Configure the Taxonomy Module
This section describes how to set up the HPE ICC modules, and discusses the following
configurations:
•
•
•
•
•
•
Enable Automatic Query Guidance (AQG)
18
Configure Document Viewing
18
Enable Secure Connections
19
Enable Encrypted HPE IDOL Connections
19
Enable Paging Managers
20
Enable Binary Categories
20
If you are using the Applications module, see Activate the Applications Module, on page 130.
Enable Automatic Query Guidance (AQG)
Enable AQG in the HPE IDOL Server to receive guidance on how to build query text for categorization
or training purposes. If VQL is enabled, you can also receive guidance on how to build query text for
topic management.
To enable AQG
1. Open the HPE IDOL server configuration file in a text editor.
2. In the [Server] section, set the following parameters:
QuerySummaryAdvanced=True
QuerySummaryIDs=True
3. Set QuerySummaryLength to 25.
4. Set QuerySummaryTerms to 50.
5. Restart the HPE IDOL server.
Configure Document Viewing
You can configure HPE IDOL Collaborative Classifier for extended document viewing. This allows you
to launch a three-pane view of a particular document from any of the review document lists.
To configure HPE ICC for extended document viewing
1. Go to <installDir>/IDOLserver/<hostname>/templates.
2. Open the RetrievalAppConfig.tmpl file. Set the idolHost and idolPort settings to the HPE
IDOL Server’s Host and ACI Port.
3. Restart HPE IDOL Server.
4. Restart HPE ICC.
HPE IDOL Collaborative Classifier (11.1)
Page 18 of 155
User Guide
Chapter 2: Installation and Configuration
Enable Secure Connections
You can enable secure connections between HPE IDOL Collaborative Classifier and HPE IDOL
Server through the HTTPS protocol. Consider secure connections if enhanced communication security
is important.
To enable secure connections
1. Configure Secure Socket Layer (SSL) support in HPE IDOL Server. See the HPE IDOL Server
Administration Guide.
2. Restart HPE IDOL Server.
3. On the HPE ICC bootstrap page, select the HTTPS protocol option.
4. Complete the bootstrap procedure.
Enable Encrypted HPE IDOL Connections
Enable the encrypted HPE IDOL connections setting to encrypt every call from the HPE ICC server to
HPE IDOL Server.
NOTE:
To use encrypted connections, related HPE IDOL components must contain a valid OEM
license key file.
With encryption enabled, when you double-click a document in any document table, HPE ICC passes
the XML content of the document from HPE IDOL to the client machines as a temporary XML file. The
browser on the client machine is used to display the temporary XML file.
NOTE:
To display document content by using encrypted connections, you must set a browser as the
default application to open XML file.
To use encrypted connections from bootstrap
l
The HPE ICC bootstrap page now has options to select an HPE IDOL connection type. By default,
the Unencrypted type is selected. Select Encrypted to use encrypted connections.
To use encrypted connections after bootstrap
1. Go to the TomcatX/webapps/icc/ folder, where X is the Tomcat version number.
2. Open the icc.cfg configuration file in a text editor.
3. Set IDOL_CONNECTION_TYPE to Encrypted (by default, this parameter is set to Unencrypted).
4. Save the file.
5. Restart Tomcat.
HPE IDOL Collaborative Classifier (11.1)
Page 19 of 155
User Guide
Chapter 2: Installation and Configuration
Enable Paging Managers
Enable paging managers to navigate the Document List in the Taxonomy module (see Manage
Documents, on page 105) and view more than the first 100 documents in the list.
To enable paging managers
1. Go to <installDir>\TomcatX\webapps\icc\WEB-INF\classes\config\, where X is the
Tomcat version number.
2. Open acc.properties, and set the paging configuration:
panel.documents.list.paging.enable parameter=True
3. Set the category.query.max.result parameter to the desired number of results per page. For
example:
category.query.max.result=100
NOTE:
HPE recommends that you set a value between 50 and 200.
4. Restart Tomcat.
Related Topics
l
Paging Managers, on page 109
Enable Binary Categories
To view and use the Manage Binary Categories window, you must first enable the HPE IDOL
Collaborative Classifier binary category feature.
To enable binary categories in HPE ICC
1. Go to <installDir>\TomcatX\webapps\icc\WEB-INF\classes\config, where X is the Tomcat
version number.
2. Open acc.properties, and set the category.binary.enable parameter to True.
3. Restart Tomcat.
Related Topics
l
About Binary Categories, on page 47
l
Author Binary Categories, on page 69
l
Populate Binary Categories, on page 101
l
Publish Binary Categories, on page 123
HPE IDOL Collaborative Classifier (11.1)
Page 20 of 155
User Guide
Chapter 2: Installation and Configuration
Install and Access the HPE IDOL Collaborative Classifier
Client
After the initial system configuration, new users access the Client Web Start Installer page directly.
Each client user can bookmark the URL and launch HPE ICC from their browser.
NOTE:
HPE ICC is a configurable product. All descriptions, paths, and instructions in this guide relate
to an out-of-the-box installation. Contact your administrator for possible changes made to your
installation.
To install the HPE ICC Client
1. Open your browser and enter the following URL:
http://<hostname>:<port>/icc
where hostname is the name of the host on which HPE ICC is installed and port is the port
number of the Tomcat server.
The Client Web Start Installer page opens.
2. Click the link provided to check whether Java Web Start is installed.
3. Click Launch Collaborative Classifier Client.
A progress bar is shown during the installation.
4. If you are installing HPE ICC for the first time, a Security Warning dialog box opens after the
installation is complete. Click Run to accept the certificate and continue.
5. If you are installing HPE ICC for the first time, an HPE IDOL Collaborative Classifier Desktop
Integration dialog box opens. Click Yes if you want to add an HPE IDOL Collaborative Classifier
shortcut icon to your desktop.
6. The HPE IDOL Collaborative Classifier login dialog box opens.
7. Enter your user ID and password.
8. Select a language from the Language list.
9. Click Login.
HPE ICC starts.
HPE IDOL Collaborative Classifier (11.1)
Page 21 of 155
User Guide
Chapter 2: Installation and Configuration
HPE IDOL Collaborative Classifier (11.1)
Page 22 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Chapter 3: Overview of HPE IDOL Collaborative
Classifier
This section discusses how to navigate and use the HPE ICC modules.
•
•
•
•
Icons
23
Windows
26
Menus
27
Name Restrictions
30
Icons
The following tables display and describe the common icons in the user interface.
Icon
Description
Lock category
Unlock category
Refresh
Add
Remove
Search for category node in tree
Category locked for editing (the pencil can appear on new, published, or modified categories)
New category (white: no category rule defined)
Modified category (yellow: category rule defined)
Published category (green spot)
Reference category (the reference arrow can appear on new, published, or modified
categories)
Create subtaxonomy
Export taxonomy to XML file
Published category that has been edited since publication (green spot with white center)
HPE IDOL Collaborative Classifier (11.1)
Page 23 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
A white lock on a category indicates that the category is locked by someone else.
Taxonomy Module Icons
The following table lists the icons used to manage documents in the Taxonomy module.
Icon Description
Document To Be Reviewed
Document Accepted
Excluded document
Manually included document
Automatically included document
Column header: sequence number of documents shows
Recommend a document to another category
Relevant for training
Column header: icon shows document type
Remove from manual actions
Change the rank of a manually included document
Column header: check mark icon for training documents
Categorization and QA Icons
The following table displays the icons used on the Categorization and QA tabs.
Icon
Description
Open query text or filter wizard
Test query or filter
Create query from rules of child categories
Show history of query text or filter (also reviewer notes)
HPE IDOL Collaborative Classifier (11.1)
Page 24 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Icon
Description
Relevant for training
Remove document from training or from benchmark
Input mode - toggles Query input mode
Open Automatic Query Guidance (if enabled)
Use query as a category rule
Relevant for benchmark
Manage Binary Category Icons
There are several unique icons in the Manage Binary Category window, and some icons have slightly
different meanings, as shown in the following table. For more information about Binary Categories, see
Binary Category Features, on the next page.
Icon
Description
New or modified binary category
Locked for editing
Modified after publishing
Relevant for positive training
Directory relevant for positive training
Binary category published to host server
Locked by another user
Binary Taxonomy root (this icon never changes, because you cannot publish the root)
Relevant for negative training
Directory relevant for negative training
HPE IDOL Collaborative Classifier (11.1)
Page 25 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Windows
This section discusses the common user interface elements.
HPE IDOL Collaborative Classifier includes icons, tooltips, and other navigation features.
Summary Window
When you first open HPE ICC, the Summary window is displayed. Depending on your user role and
which modules are enabled, you might see the Taxonomy module, the Application module, or both of
the modules.
NOTE:
By default, only the Taxonomy module is shown. To display the Applications module, enable it
in the acc.properties file. See Activate the Applications Module, on page 130.
The elements of the Summary window include:
l
The File, Window, and Help menus
l
The module panes
l
The Refresh button, which refreshes the module information
l
User and server information, displayed in the bottom bar
Interface Panes
The HPE ICC interface is divided into multiple panes. For example, the draft taxonomy view displays
the following panes:
l
The Taxonomy pane
l
The Resource Taxonomies pane (drag the bottom bar up to reveal this pane)
l
The working area pane, including the Documents, Category Details, Categorization, QA, and
Permissions tabs
Drag-and-drop functionality is available in the Taxonomy pane, as well as from the resource taxonomy
pane to the taxonomy pane.
Binary Category Features
When enabled, you can manage binary categories through the Manage Binary Categories window,
which is shown when you open the Binary Taxonomy project from the Summary window.
The Manage Binary Categories window is separate from the regular taxonomy window, so both can
be open at the same time. The layout, icons, and functionality are similar to those of regular
taxonomies, with several important differences. See Author Binary Categories, on page 69 for a more
detailed explanation.
HPE IDOL Collaborative Classifier (11.1)
Page 26 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Related Topics
l
Enable Binary Categories, on page 20
l
About Binary Categories, on page 47
l
Author Binary Categories, on page 69
l
Publish Binary Categories, on page 123
Menus
HPE IDOL Collaborative Classifier menus are described in the following sections. The Manage Binary
Categories window does not contain a menu bar.
NOTE:
Depending on your user role, you might not have access to all menu options.
The File, Window and Help menus are available in both HPE ICC modules.
File Menu
The File menu choices include:
l
New. Create a new Taxonomy (or application, if enabled).
l
Open. Open a Taxonomy (or application, if enabled).
l
l
Administration. Open the administration interface. See HPE IDOL Collaborative Classifier
Administration, on page 33.
Preferences. Open the Preferences dialog box.
To reset preferences to the default settings
1. In the File menu, select Preferences....
The Preferences to Reset dialog box opens.
2. Select or clear the preferences that you want to reset. Click OK.
HPE ICC resets the selected preferences to the default settings.
Window Menu
The Window menu lists the open windows in HPE ICC. Use this menu to select the window that you
want to display. An example of the window menu might show the Summary window and an open
taxonomy.
Help Menu
The Help menu contains two items:
About
Display the version information.
Documentation
Displays this guide online.
HPE IDOL Collaborative Classifier (11.1)
Page 27 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
View Menu
The View menu controls which view you are working in for the module.
NOTE:
The views available to you are based on the module and your user role.
The different views enable you to see a resource in particular states. The following views are available:
Draft
View
The main working view. This view displays how the module might appear after it is
complete, approved, and published.
Filtered
Draft
View
A draft view, filtered by preset or user-defined filters in the Filter menu. To display the
corresponding categories, select Filtered Draft View, and then select a filter (this
feature works only in the Taxonomy module).
Published This view displays the module’s published nodes. Publishers can use this view to see
View
previously published nodes.
Taxonomy Filters Menu
The Filters menu controls filtered views. You can set filters to view documents that meet certain
requirements.
HPE ICC contains the following predefined filters:
To be Reviewed
Show categories with documents to be reviewed, and that can be edited by the
user.
Workspace
Show modified categories that are locked and editable.
Publish
Requests
Show modified categories that are locked and publishable.
Locked
Show locked categories.
Editable
Show editable categories.
Filters require an explicitly defined user role. If administrators need to use filters, add them to the
taxonomy as publishers.
For information about using filters and creating custom filters, see About Filters, on page 67.
Resources Menu
The Resources menu enables you to open and close resource taxonomies.
Module Navigation Shortcut Menu
Right-click a tree node in the module navigation pane to access the Shortcut menu.
HPE IDOL Collaborative Classifier (11.1)
Page 28 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
NOTE:
The actions available to you in the shortcut menu are based on your user role.
Refresh
Refresh the subcategories in the category.
Publish or Request
Publish
Publish or request publication of the category.
Edit
Lock the category to enable editing.
Release Edit Lock
Unlock the category.
Acquire Subtree Edit
Locks
Lock all the categories in the subtree for which you have permission, and
that are not locked by other users.
Release Subtree Edit
Locks
Unlock the category and its subtree.
Lock All Categories
Lock all categories in the taxonomy.
Unlock All
Categories
Unlock all categories in the taxonomy.
Add Child
Add a child node to the category.
Copy
Copy a category or a subtree.
Paste as Reference
Create a reference category.
Paste as Copy
Paste a copied subtree.
Mark for Deletion
Mark the category and its subcategories for deletion. To delete, publish the
marked categories.
Restore
Restore the deleted category. This action applies to individual categories
only.
Find
Find a tree node.
Rename
Rename the category.
Show References or
Show Primary
Category
Display the references of a specific category, or select the primary
category of a reference category. The categories are displayed in a filtered
draft view.
Taxonomy Filtered Draft Shortcut Menu
Options in the category pane menu change if you are using filters. Use the toolbar or right-click when
your cursor is on a category to access the filtered draft menu. You must first apply a filter to view this
menu.
NOTE:
The actions available to you in the taxonomy filtered draft menu are based on your category
privileges.
HPE IDOL Collaborative Classifier (11.1)
Page 29 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Refresh
Refresh the category and its subcategories.
Edit
Lock the category to enable editing.
Release Edit Lock
Unlock the category.
Show in Draft View Display the category in draft view.
Publish
Publish or request publication of the category.
Published Shortcut Taxonomy Menu
Options in the category pane menu change if you are in a published view. Right-click when your cursor
is on a category in the categories pane to perform the Refresh action.
Query Tree Shortcut Taxonomy Menu
A shortcut menu is available in the Taxonomy module when you are editing a query tree.
NOTE:
The actions available to you in the query tree shortcut taxonomy menu are based on your
category privileges.
Add Child
Add a new node to the tree.
Edit Node
Edit the selected node name.
Remove Subtree
Delete the selected subtree in the query tree.
Copy Subtree
Copy the selected subtree.
Paste Subtree
Paste the copied subtree into the location selected in the query tree.
Expand Subtree
Expand a collapsed subtree.
Collapse Subtree Collapse an expanded subtree.
Name Restrictions
The following sections describe the rules for naming resources in the taxonomy and application
modules. For more information about resourses, see About Resource Administration, on page 33.
Taxonomy Module Resource Name Restrictions
You cannot use the following characters in resource names.
\ backslash
/ forward slash
HPE IDOL Collaborative Classifier (11.1)
Page 30 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
Application Module Resource Name Restrictions
You cannot use the following characters in resource names.
\
backslash
/
forward slash
,
comma
:
colon
*
asterisk
@ at sign
?
question mark
“
double quotation mark
<
left angle bracket
>
right angle bracket
|
pipe
HPE IDOL Collaborative Classifier (11.1)
Page 31 of 155
User Guide
Chapter 3: Overview of HPE IDOL Collaborative Classifier
HPE IDOL Collaborative Classifier (11.1)
Page 32 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
Chapter 4: HPE IDOL Collaborative Classifier
Administration
This section describes how to administer the HPE ICC modules.
•
•
•
•
•
About Resource Administration
33
Maintain Modules
33
Manage Users
34
Change Default Taxonomy Properties
36
Edit an Application URL
37
About Resource Administration
To access the administration features, you must have administrator privileges. See User Permissions,
on the next page.
To open the Administration menu
l
In the File menu, select Administration.
The Administration window opens.
The Administration window consists of the following elements:
l
Resources Displays the list of available taxonomies and applications. The pane content is
controlled through the Show list.
l
Actions. Displays the possible actions for the selected resource.
l
Properties. Displays the default properties of the selected list item.
Maintain Modules
You can use the Administration screen to create and delete resources.
NOTE:
The applications module is disabled by default. To use the applications module and application
resources, see Activate the Applications Module, on page 130.
Create a Resource
To add a new resource
1. In the File menu, select Administration.
2. In the Resources pane, select the resource type (for example, Taxonomy) from the Show list.
HPE IDOL Collaborative Classifier (11.1)
Page 33 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
3. In the Actions pane, click Add New (Resource Type).
The applicable wizard opens. The Taxonomy module wizard is described in Add Taxonomies, on
page 51. The Application module wizards are described in Add Applications, on page 131.
Delete a Resource
To delete a resource
1. In the File menu, select Administration.
2. In the Resources pane, select the resource type (for example, Taxonomy) from the Show list.
3. Select the resource you want to delete from the list in the Resource pane.
4. Click Remove (Resource Type) under Actions.
A confirmation dialog box opens. To remove the taxonomy from the published taxonomy server,
select Also remove from published taxonomy server.
5. Click OK to remove a taxonomy, or Yes to remove an application.
Manage Users
You can manage users from the Administration screen. You must add each user to the user list. After
you add a user, you can assign permissions for the modules to each user. For example, you can add a
user named jane_doe and give her permission to open applications and create taxonomies. Jane Doe
can see the Application and Taxonomy modules in the Summary window when she logs into HPE ICC.
Jane can only open applications, but can open and create taxonomies.
Users with Open privileges for the Taxonomy module can open and view items. Until the user is
assigned additional privileges (reviewer, editor, or publisher) for an individual resource, they can only
view resources.
Users with Create privileges in the Taxonomy module can create new resources. They are
automatically assigned publisher roles for the resources that they create.
The Application module does not have additional user-role permissions for each application; users with
Open privileges can edit and publish.
User Permissions
The HPE IDOL Collaborative Classifier modules have the following permission levels:
l
Administrator access. Only HPE IDOL administrators have administrator access to the modules.
l
Module access. Administrators can assign access and creation permissions for each module.
l
Role-based workflow access. Reviewer, Editor, and Publisher permissions are assigned based on
user roles. See Module User Roles, on the next page for information about the user roles available in
the different modules.
HPE IDOL Collaborative Classifier (11.1)
Page 34 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
Module User Roles
You can control functionality in the HPE ICC modules by assigning a user role to each module user.
The following table lists the roles available and the tasks that each role can access.
Reviewer
The reviewer can open a module, and can view that module’s resources. Reviewers
can modify reviewer notes in the Taxonomy module.
Editor
The editor can modify taxonomy structures (for example, add or edit category
nodes). The editor can also edit the node’s category details, such as user-defined
properties. The editor can use all actions in the Draft View except for Publish.
Publisher
A publisher can perform the same tasks as an editor, and can also publish category
nodes.
Administrator The administrator can perform all the actions of the publisher, and has access to the
Administration tasks, such as adding, maintaining, and removing resources. For
more information, see HPE IDOL Collaborative Classifier Administration, on page
33.
NOTE:
The administrator role is not assigned through HPE ICC. Administrator tasks are performed by
the administrator user type (see User Permissions, on the previous page).
Taxonomy module user roles
Role
Review Nodes Edit Nodes Publish Manage Resources
Reviewer
Yes
Editor
Yes
Yes
Publisher
Yes
Yes
Yes
Administrator Yes
Yes
Yes
Yes
NOTE:
These roles are available only in the Taxonomy module. The Application module has only two
user roles, the administrator and the user who maintains the applications. Both these roles can
edit and publish.
Add New Users
To add a new user
1. From the File menu, select Administration.
2. From the Users menu, select Manage.
3. The Manage Permissions dialog box opens.
4. Click +.
HPE IDOL Collaborative Classifier (11.1)
Page 35 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
A new user dialog box opens.
5. Enter the user name and password. Click OK.
The new user is displayed in the Available Users list.
Manage User Permissions
To manage user permissions
1. From the File menu, select Administration.
2. From the Users menu, select Manage.
The Manage Permissions dialog box opens.
3. From the Available Users list, select the user whose permissions you want to modify, and then
click the up arrow.
The user is displayed in the Existing Users list.
4. Select the Open or Create check box. If you select Create, Open is automatically selected and
cannot be deselected.
5. Click Save Changes.
Delete a User
To delete a user
1. From the File menu, select Administration.
2. From the Users menu, select Manage.
The Manage Permissions dialog box opens.
3. In the Available Users list, select the user you want to remove and click -.
The user is removed.
4. Click OK.
Change Default Taxonomy Properties
The default properties determine the starting thresholds and HPE IDOL properties of the selected
taxonomy and its children. You define default properties when you create the taxonomy (see Add
Taxonomies, on page 51).
NOTE:
Changing a taxonomy’s default properties does not change its current property values; only
subsequently added children inherit the default property values.
To change the default properties
1. From the File menu, select Administration.
2. In the Resources pane, select the taxonomy that you want to edit.
HPE IDOL Collaborative Classifier (11.1)
Page 36 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
The Properties dialog box opens, listing the default property values.
NOTE:
You can change the default Threshold Accepted and Threshold Rejected values as well as
the HPE IDOL NumResults value. For more information on thresholds, see About the
Document Distribution Graph, on page 109, and for more information about HPE IDOL
properties, see Edit HPE IDOL Standard Properties, on page 65.
3. Change the desired property value or values.
4. Click Save Modifications.
The new property values take effect with any new child taxonomy.
Edit an Application URL
You can use the application module to edit web server applications. The application URL is the URL
that the browser is pointed to in order to use the application itself.
NOTE:
The Application Module is disabled by default. To enable the module, see Activate the
Applications Module, on page 130.
To add or update a URL for an application
1. From the File menu, select Administration.
2. In the Resources pane, select Applications from the Show list.
3. Select the application that you want to edit from the list in the Resource pane.
4. In the Properties section, enter or edit the URL for your application in the View URL text box.
5. Click Save Modifications.
HPE IDOL Collaborative Classifier (11.1)
Page 37 of 155
User Guide
Chapter 4: HPE IDOL Collaborative Classifier Administration
HPE IDOL Collaborative Classifier (11.1)
Page 38 of 155
User Guide
Part 2: Taxonomies
This section describes how to create and use taxonomies.
l
About Taxonomies
l
Create Taxonomies
l
Author Taxonomies
l
Author Binary Categories
l
Populate Categories
l
Populate Binary Categories
l
Manage Documents
l
Review Categories
l
Publish and Export a Taxonomy
HPE IDOL Collaborative Classifier (11.1)
Page 39 of 155
User Guide
HPE IDOL Collaborative Classifier (11.1)
Page 40 of 155
User Guide
Chapter 5: About Taxonomies
Chapter 5: About Taxonomies
This section discusses the concepts of the Taxonomy module.
•
•
•
•
•
•
•
Classify Information
41
Create, Populate, and Use Taxonomies
42
•
•
•
•
Build the Taxonomy
42
Define Categories
44
Populate the Taxonomy With Documents
44
Use the Taxonomy
45
About Category Rules
45
About Categories and Taxonomies
46
•
•
•
Categories and Properties
46
Reference Categories
46
Resource Taxonomies
47
How the Taxonomy Module Handles Documents
47
About Databases
47
About Binary Categories
47
Classify Information
Classifying information into categories and subcategories allows your users to locate individual
documents. A user would descend through the concepts represented by each category until they find
relevant documents. Classification organizes the documents in your collections into content
hierarchies, called taxonomies. A taxonomy is a hierarchical organization of information in categories,
such as types of animal, as shown in Figure 1: Sample classification containing animal categories,
below.
Figure 1: Sample classification containing animal categories
HPE IDOL Collaborative Classifier (11.1)
Page 41 of 155
User Guide
Chapter 5: About Taxonomies
When the number of documents used by an organization increases, the need to classify them using an
intuitive and meaningful hierarchy becomes more important.
A major issue for many organizations is creating a hierarchy without examining each document,
especially when the categories themselves have not already been explicitly identified. This process
becomes more difficult when there are thousands or millions of documents to classify.
HPE’s classification solution enables a flexible approach that can combine subject matter expertise
with automatic classification for documents in all popular formats. This solution enables you to
organize your information assets into categories, making them accessible for users to browse.
This section describes how to classify information using the Taxonomy module.
Create, Populate, and Use Taxonomies
The basic stages are:
1. Build the Taxonomy
2. Define Categories
3. Populate the Taxonomy With Documents
4. Use the Taxonomy
NOTE:
In some situations you can combine the first two stages, or even the first three stages. These
situations are also identified in the following sections.
Build the Taxonomy
You can define a view of your content by organizing information into a hierarchy of categories (a
taxonomy). The result is a skeletal navigation structure of categories, each with a name. Figure 2:
Sample taxonomy navigation containing animal categories, on the next page shows an Animals
category with child categories labeled Cats and Birds. These child categories can have their own
children, and so on.
HPE IDOL Collaborative Classifier (11.1)
Page 42 of 155
User Guide
Chapter 5: About Taxonomies
Figure 2: Sample taxonomy navigation containing animal categories
Several techniques exist for building a taxonomy:
l
l
l
Use domain experts. Taxonomies built by domain experts are quite common. A domain expert
typically builds a skeleton taxonomy and assigns names to its categories.
Import from a hierarchy. Users can create a taxonomy by extracting the implicit hierarchies from
URLs or file system hierarchies, or hierarchies defined in metadata (such as the Dewey decimal
number in a library catalog). Users can then mirror these hierarchies in a taxonomy.
Use HPE taxonomy packages. HPE provides taxonomy packages in a number of industries to
help companies start their categorization initiatives. Other vendors also provide taxonomies for
particular industry segments.
To purchase HPE taxonomy packages, contact your HPE representative.
l
l
Convert third-party taxonomy packages. To use a third-party taxonomy, you must convert your
taxonomy into an XML format. Use the taxonomy.dtd to define the XML format. For more
information, see HPE Category XML Format, on page 145.
Use concept extraction and naming. The HPE classification engine automatically extracts key
concepts contained in a set of documents and organizes them into a hierarchy. The HPE engine
analyzes your documents and groups concepts that recur throughout the content into categories.
The engine then automatically creates a taxonomy structure for these concepts. Automatic naming
uses linguistic analysis to generate labels for these categories.
See Automatically Create a Taxonomy from a Database, on page 53.
HPE IDOL Collaborative Classifier (11.1)
Page 43 of 155
User Guide
Chapter 5: About Taxonomies
Define Categories
After you create the taxonomy, the next step is to attach category definitions to the taxonomy so that
you can populate the taxonomy with documents. Each category definition consists of a mathematical
rule against which each document can be evaluated for membership in that category. A simple (and
incomplete) rule for the category Animal could be defined as follows:
If a document contains the word ‘paw’ or ‘hoof,’ include the document in the Animal category.
There are several techniques for building rules that define categories:
l
l
l
l
Expert-defined rules. A domain expert can define a category rule for each category. These rules
are sometimes called business rules because the domain expert typically tailors them to be relevant
to a specific business or business function.
Import from a hierarchy. You can define categories by extracting the implicit hierarchies from
URLs or file system hierarchies, or hierarchies defined in metadata (such as the Dewey decimal
number in a library catalog). This technique is useful if membership in categories corresponds to an
implicit hierarchy, such as a file system or URL hierarchy. In this situation, the first two stages,
building the taxonomy and defining categories, are combined into one stage.
Use industry-standard categories. A standards body or independent vendor can create category
definition rules for an industry’s vertical taxonomies. This technique is closely associated with using
industry-specific taxonomies. You can combine the use of industry-standard taxonomies with
standard categories.
Create automatic categories. You can supply example documents known as training documents
to an automatic classification system that creates categories. The system learns from these training
documents and creates a defining rule for each category.
HPE IDOL implements automatic category creation.
l
Create automatically using a classification engine. When you use the HPE classification engine
to create a taxonomy or subtaxonomy, each category in the taxonomy or subtaxonomy is
associated with a list of concept terms (nouns and noun phrases). You can use these concept terms
to automatically create a category rule for the category.
With the HPE classification infrastructure, you can implement any combination of these methods. Best
results are often achieved by combining automatic capabilities with domain expertise.
Populate the Taxonomy With Documents
After you have built the taxonomy and associated a rule with each of its categories, you can populate
the categories with documents. You can populate the taxonomy either manually or automatically, or in
both ways:
l
l
Custom. For each document, an expert determines the categories to populate, and then explicitly
populates those categories in the taxonomy with the document.
Automatic. The system evaluates each document against the rule for each category and assigns
the document to the appropriate categories in the taxonomy.
You can combine both methods. You can assign documents to categories based on defined rules, or
you can manually override those rules and place a document in a different category or in additional
categories.
HPE IDOL Collaborative Classifier (11.1)
Page 44 of 155
User Guide
Chapter 5: About Taxonomies
Use the Taxonomy
After building the taxonomy, defining categories, and populating the taxonomy, your taxonomy is ready
to use. With well-implemented taxonomies, users perform a balanced combination of search and
browsing to locate documents. The search could be issued at the top level to filter the categories in
which matching documents exist, or contained in a subset of the taxonomy.
You can create multiple taxonomies to organize content in ways that make the most sense for each
group in your enterprise. For example, you can create separate taxonomies for the sales, marketing,
human resources, and engineering departments.
About Category Rules
Category rules are fundamental building blocks in the HPE classification infrastructure. Category rules
play a central role in the category definition stage, described in Define Categories, on the previous
page.
A category rule is a stored query expression. A category rule models a concept of interest, which is
used as the definition for a category.
When a category rule is evaluated against a set of documents, the search engine identifies the subset
of the documents that match the concept that the category rule represents.
By using individual query expressions, or by combining them, you can create category rules that you
can use to decide whether a document belongs to the category. There are several techniques for
constructing category rules, ranging from domain expertise to the use of automated machine-learning
techniques.
You can combine category rules regardless of how they are created. One advantage of combining
query expressions is that it allows the gradual construction of complex category rules in such a way
that you can share basic query expressions between multiple higher-level category rules.
Consider the following scenario:
l
You can use the expression GM <OR> Ford <OR> Chrysler to model the concept "North American
car manufacturers."
When you use this expression to evaluate a set of newspaper articles, the search engine selects all
articles that mention GM, Ford, or Chrysler as matching the concept "North American car
manufacturers."
l
l
You can combine category rules with Boolean operators to create more complex category
definitions. For example, you might combine the concept “North American car manufacturers” with
“European car manufacturers.” By combining these and applying <NOT> to the concepts, you could
create a new category definition corresponding to the concept “Asian car manufacturers” (this
definition assumes no South American or Australian car manufacturers).
You can also use sophisticated non-Boolean operators.
HPE IDOL Collaborative Classifier (11.1)
Page 45 of 155
User Guide
Chapter 5: About Taxonomies
About Categories and Taxonomies
Every taxonomy has a hierarchical structure with a single top-level root category. All category nodes
have at least one parent category and can have zero, one, or more children (subcategories).
Figure 3: Taxonomy terminology
Categories and Properties
Categories in the Taxonomy module are the nodes in the taxonomy that content is added to, according
to a specific category rule. Category names identify each category node in the taxonomy.
The category name is one of the category properties. Another property is the optional description that is
added in the Category Details tab in the HPE IDOL Collaborative Classifier Taxonomy module.
Reference Categories
Sometimes a category belongs in multiple places in a taxonomy. For example, a category
corresponding to a specific product on a company's intranet might belong under both the Research and
Development category and the Marketing category. Rather than duplicating the category and all its
documents, you can copy the category and create a reference from the second category to the original.
The two categories now share the same properties and document lists, but reside in different parent
categories.
Reference categories can also have different user assignments.
Reference categories behave similarly to UNIX file system symbolic links. This type of category can
appear in multiple locations in a taxonomy. Reference categories are categories that inherit the
properties and document lists from their referred primary categories.
l
Reference categories cannot have child categories.
l
Reference categories cannot refer to siblings.
l
You cannot publish reference categories unless you also publish their primary categories.
HPE IDOL Collaborative Classifier (11.1)
Page 46 of 155
User Guide
Chapter 5: About Taxonomies
Resource Taxonomies
In the Taxonomy module, you can use resource taxonomies to copy categories into a draft taxonomy.
Resource taxonomies include published taxonomies.
You can use multiple taxonomies as resource taxonomies. Resource taxonomies are read-only, but
you can use drag-and-drop functionality to copy categories directly into your draft taxonomy.
How the Taxonomy Module Handles Documents
Documents from associated collections are available for users to create category content. This content
is shown in the document list of the Documents tab in a taxonomy. You can add documents for this list
manually or automatically depending on the thresholds set in the document distribution graph.
In the draft view of the Taxonomy module, you can manipulate documents until you are satisfied with
the document list for a category.
All documents to be published must be in the Accepted area. Place documents that require additional
review in the To Be Reviewed area. You can remove documents that you do not want in the category
from the view. This manual manipulation of the documents assists the category editor in creating the
exact document set required for the category.
Publishing the category publishes all documents in the Accepted area and makes those documents
available for your users. You can add additional documents manually to update published categories.
For information about views and user roles, see Overview of HPE IDOL Collaborative Classifier, on
page 23.
About Databases
The Taxonomy module associates databases with taxonomies to help users to create different
document lists for different databases for each category.
The Taxonomy module can use the databases that exist on the training HPE IDOL server and that
show in the Database list.
About Binary Categories
Binary categories are used for binary queries (see Binary Queries, on page 103). Unlike regular
categories, which receive only positive training, binary categories can receive both positive and
negative training. For more information on binary categories, see the HPE IDOL Server Administration
Guide.
In the context of HPE IDOL Collaborative Classifier, you can manage binary categories through the
Manage Binary Categories window. You must enable this feature in the acc.properties
configuration file. You can use the Manage Binary Categories window to set positive and negative
training documents, text, and directories.
HPE IDOL Collaborative Classifier (11.1)
Page 47 of 155
User Guide
Chapter 5: About Taxonomies
Related Topics
l
Enable Binary Categories
l
Author Binary Categories, on page 69
l
Populate Binary Categories, on page 101
l
Publish Binary Categories, on page 123
HPE IDOL Collaborative Classifier (11.1)
Page 48 of 155
User Guide
Chapter 6: Create Taxonomies
Chapter 6: Create Taxonomies
This section describes how to add taxonomies in the Taxonomy module.
•
•
Work with Taxonomies
49
Add Taxonomies
51
For more information about taxonomies, see About Taxonomies, on page 41.
Work with Taxonomies
You can use the Taxonomy module to create and maintain taxonomies and categories. You can then
publish the taxonomies to create searchable content. This section describes the basic procedures for
working with taxonomies.
•
•
•
•
Open and Close Taxonomies
49
Select a Taxonomy View
50
Edit Categories
50
Request Publication of a Category
50
Open and Close Taxonomies
You can open and close taxonomies in the draft and published views. You can also switch between
open taxonomies through the Window menu.
To open a Taxonomy from the menu
1. In the File menu, select Open, and then select Taxonomy.
The Open Taxonomy dialog box opens.
2. Select a taxonomy from the list.
3. Click OK.
The taxonomy opens.
To open a Taxonomy from the Taxonomy Summary Window
1. In the Summary window, select a taxonomy from the Taxonomy pane.
2. Click Open.
The taxonomy opens.
NOTE:
You can also double-click the taxonomy to open it.
To close a taxonomy
l
In the File menu, select Close.
HPE IDOL Collaborative Classifier (11.1)
Page 49 of 155
User Guide
Chapter 6: Create Taxonomies
Opening and Closing Resource Taxonomies
When creating a taxonomy, you might want to use the category rule created for another category in
another taxonomy.
You can open and close resource taxonomies in the draft view. You can open more than one at a time,
but only one is displayed at a time in the resource taxonomy pane.
To open a Resource Taxonomy
1. In the Resources menu, select Open Resource Taxonomy.
The Open Resource Taxonomy dialog box appears.
2. Select a taxonomy from the list. Click OK.
The resource taxonomy appears in the list of resource taxonomies in the resource taxonomy pane.
To close a Resource Taxonomy
l
In the Resources menu, select Close Resource Taxonomy.
Select a Taxonomy View
You can view draft taxonomies, draft taxonomies with filters applied, and published taxonomies. To
change views, select a view from the View menu.
Edit Categories
To edit information in a category, you must lock the category. To release the category for other users to
edit that category, you must unlock the category.
NOTE:
Only an administrator can unlock categories locked by other users.
To lock a category for editing
1. In the Taxonomy pane, select the category.
2. Click Edit
.
To unlock a Category
1. In the Taxonomy pane, select the category.
2. Click Release Edit Lock
.
TIP:
You can unlock entire subtrees with the right-click menu available in the taxonomy pane.
Request Publication of a Category
If you have Editor privileges, you can request publication of a category.
HPE IDOL Collaborative Classifier (11.1)
Page 50 of 155
User Guide
Chapter 6: Create Taxonomies
NOTE:
If you have Reviewer privileges only, you cannot request publication of a category. See Publish
and Export a Taxonomy, on page 121 for information about how to publish.
To request publication of a category
1. Select the category in the taxonomy pane.
2. Click Request to Publish.
If there is a subtree of child categories under the category node, a confirmation dialog box opens.
3. Select This node only to publish a single node or Entire subtree to publish the entire subtree of
categories.
4. Click OK.
If there are no child categories, a dialog box opens asking you to confirm the request for
publication.
5. Click Yes.
HPE ICC sends a request to the category Publisher. The lock transfers to the publisher.
Add Taxonomies
You can use the taxonomy wizard to create new taxonomies.
•
•
•
Create an Empty Taxonomy
51
Import a Taxonomy
52
Automatically Create a Taxonomy from a Database
53
Users must have Create privileges to create taxonomies. Users are automatically assigned Publisher
roles for any taxonomies they create. For more information, see Manage Users, on page 34.
Create an Empty Taxonomy
To create an empty taxonomy manually, use the Add New Taxonomy wizard.
To create an empty taxonomy manually
1. From the Summary window, click Create in the Taxonomy pane.
The New Taxonomy dialog box opens.
2. Select Create a taxonomy manually.
3. (Optional) Change the taxonomy default properties. You can change the default properties after
the taxonomy is created through the Administration interface. See Change Default Taxonomy
Properties, on page 36.
4. Click Next.
The Create Taxonomy Manually dialog box opens.
By default, Create empty taxonomy is selected.
5. Enter the name of the new taxonomy. Enter the name and port of the HPE IDOL server or HPE
HPE IDOL Collaborative Classifier (11.1)
Page 51 of 155
User Guide
Chapter 6: Create Taxonomies
Category server to which you want to publish the taxonomy.
Click Verify to verify the server connection. If the connection is valid, you can continue. If the
connection is not valid, enter a valid name and port.
6. (Optional) Choose Create from existing categories.
a. If you choose this option, click Browse.
A Select Category dialog box opens.
b. Select an available category from the dialog box.
c. Click OK.
NOTE:
If you create a taxonomy from an existing category, you can modify its children only if
you meet one or more of the following criteria:
l
You are the HPE ICC administrator
l
You are the original publisher of the source category
l
You are assigned editor permissions
7. In the Set Language and Encode panel, select the taxonomy language and character encoding.
NOTE:
The available character encodings depend on the selected language. You can change the
language and character encoding after the taxonomy is created. See Change the Category
Language, on page 62.
8. Click Finish.
The Confirm actions dialog box opens.
9. Review the taxonomy selections. Click Create Taxonomy to create the new taxonomy.
10. After the taxonomy is created, click Open New Taxonomy to open the taxonomy, or Exit to leave
the taxonomy wizard.
Import a Taxonomy
NOTE:
You can import only HPE IDOL-compliant or HPE IDOL-compatible XML taxonomies into HPE
ICC. See HPE Category XML Format, on page 145.
To import a taxonomy
1. From the Summary window, click Create in the Taxonomy pane.
The New Taxonomy dialog box opens.
2. Click Import a Taxonomy XML file stored locally.
3. (Optional) Change the taxonomy default properties. You can change the default properties after
the taxonomy is created through the Administration interface. See Change Default Taxonomy
Properties, on page 36.
4. Click Next.
The Taxonomy Alias dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 52 of 155
User Guide
Chapter 6: Create Taxonomies
5. Enter the name of the new taxonomy.
6. Enter the name and port of the HPE IDOL server or HPE Category server to which you want to
publish the taxonomy.
7. In the Set Language and Encode panel, select the taxonomy language and character encoding.
NOTE:
The available character encodings depend on the selected language. You can change the
language and character encoding after the taxonomy is created. See Change the Category
Language, on page 62.
8. Click Next.
The Choose Local TaxDef XML file to import dialog box opens.
NOTE:
You do not need to verify the HPE IDOL server connection. The check is done implicitly. If
the connection is not valid, a message is displayed.
9. Select the XML file to import. Click Finish.
The Confirm actions dialog box opens.
10. Review the taxonomy selections. Click Create Taxonomy to create the new taxonomy.
The taxonomy creation runs in the wizard.
11. When the taxonomy has been created, click Open New Taxonomy to open the taxonomy, or Exit
to leave the taxonomy wizard.
Automatically Create a Taxonomy from a Database
The Add New Taxonomy wizard allows you to automatically generate a taxonomy based on a selected
database, field, and other options.
NOTE:
The combination of selected database, field, and optional parameters that you use must be
valid.
You must know how the field is populated in the database before you define your settings for
auto-creation.
You can automatically create a taxonomy from a database in two ways:
l
l
Based on a Specific Field Value. You select a specific field value and generate a category
hierarchy. There is an option to generate category rules, which are based on the training documents
assigned automatically to categories. HPE ICC also generates terms and weights.
Based on File Path of Documents. You select a specific field value and generate a category
hierarchy. HPE ICC creates the category rules from the values found for the selected document
field:
Text=*
FieldText=MATCH{....}:FIELD_NAME
HPE IDOL Collaborative Classifier (11.1)
Page 53 of 155
User Guide
Chapter 6: Create Taxonomies
NOTE:
Because HPE ICC generates the category rules from the values of the specified document
field, it is important that you select the correct field when you use this method.
To create a taxonomy automatically from a database
1. From the Summary window, click Create in the Taxonomy pane.
The New Taxonomy dialog box opens.
2. Click Automatically create a Taxonomy from a Database.
3. (Optional) Change the taxonomy default properties. You can change the default properties after
the taxonomy is created through the Administration interface. See Change Default Taxonomy
Properties, on page 36.
4. Click Next.
The Choose name for Taxonomy dialog box opens.
5. Enter the name of the new taxonomy.
6. Enter the name and port of the HPE IDOL server or HPE Category server to which you want to
publish the taxonomy.
7. In the Set Language and Encode panel, select the taxonomy language and character encoding.
NOTE:
The available character encodings depend on the selected language. You can change the
language and character encoding after the taxonomy is created. See Change the Category
Language, on page 62.
8. Click Next.
The Choose Database to use for Automatic Creation dialog box opens.
NOTE:
You do not need to verify the HPE IDOL server connection. The check is done implicitly. If
the connection is not valid, a message is displayed.
9. Select a database from which to import fields. Click Next.
The Choose method for Automatic Creation dialog box opens.
10. Select either Based on a Specific Field value or Based on the File Path of Documents.
11. (Optional) Select or enter any Optional Settings. See Automatically Create a Taxonomy from a
Database, on the previous page for more information.
NOTE:
For the Based on the File Path of Documents method, the Assign Documents and
Generate Rule options are not available.
12. Click Next.
The Choose Field dialog box opens.
13. Select the field from which to create categories.
14. (Optional) Specify a field or hierarchy delimiter. Click Finish.
The Confirm actions dialog box opens.
15. Review the taxonomy selections. Click Create New Taxonomy to create the new taxonomy.
HPE IDOL Collaborative Classifier (11.1)
Page 54 of 155
User Guide
Chapter 6: Create Taxonomies
16. After the taxonomy is created, click Open New Taxonomy to open the taxonomy, or Exit to leave
the taxonomy wizard.
Optional Settings for Automatic Taxonomy Creation
You can modify the automatic generation process to ensure a more accurate taxonomy.
Prefix to Ignore
Enter a string to ignore in the field value. For example, you could create a taxonomy by using the
Document Reference field for URLs. In this case, if you enter ‘http:’, HPE ICC ignores this string at
the beginning of each path.
Ignore Suffix
Select this option to ignore suffixes in category names. For example, you could create a taxonomy by
using the Document Reference field for URLs:
http://www.hpe.com/products/prod1.html
http://www.hpe.com/products/prod2.html
http://www.hpe.com/products/prod3.html
http://www.hpe.com/people/person1.html
http://www.hpe.com/people/person2.html
l
l
If you select Ignore Suffix, HPE ICC creates only two categories in the taxonomy, products and
people.
If you select Ignore Suffix, HPE ICC create five categories, one for each document.
Assign Documents
Automatically assign documents to category nodes as training documents, based on the path and
hierarchy delimiter.
NOTE:
For the Based on the File Path of Documents method, Assign Documents is not available.
For example, you could create a taxonomy by using a path field, such as TAX_PATH, which has a format
that uses “/” as a delimiter, such as apple/banana/cherry. The category hierarchy becomes:
apple
banana
cherry
In this case, all documents that contain the path field apple/banana/cherry are assigned to the
cherry node as training documents.
In another example, you could create a taxonomy by using a FILE_PATH field, where each document is
of the form:
C:\test\hpe\doc1.txt
C:\test\hpe\doc2.txt
C:\test\hpe\doc3.txt
HPE IDOL Collaborative Classifier (11.1)
Page 55 of 155
User Guide
Chapter 6: Create Taxonomies
l
If you do not select Ignore Suffix, the category hierarchy becomes:
C:
test
hpe
doc1.txt
doc2.txt
doc3.txt
In this case, each docN.txt node has exactly one document assigned as a training document.
l
If you select Ignore Suffix, the category hierarchy becomes:
C:
test
hpe
In this case, all documents in the folder are assigned to the hpe node as training documents.
Generate Rule
This option is activated only if you select Assign Documents. If selected, HPE ICC generates the
category rules based on the training documents assigned automatically to categories.
NOTE:
For the Based on the File Path of Documents method, Generate Rule is not available.
HPE IDOL Collaborative Classifier (11.1)
Page 56 of 155
User Guide
Chapter 7: Author Taxonomies
Chapter 7: Author Taxonomies
This section discusses the taxonomy authoring tasks performed by users who are assigned editor
privileges in the Taxonomy module.
•
•
•
•
•
•
Author Role
57
Edit a Taxonomy
57
Edit Category Information
62
Add and Maintain Users
66
About Filters
67
View Published Categories
68
For more information about taxonomies, see About Taxonomies, on page 41 and Create Taxonomies,
on page 49.
Author Role
If you are assigned as an editor to a category in a taxonomy, you can perform all the actions that a
reviewer can perform (see Review Categories, on page 115). Also, you can:
l
Edit category node information
l
Change user roles for users with editor or reviewer privileges
l
Move documents into Accepted
l
Change the taxonomy structure
l
Use documents in a collection to train categories
l
Use topics to further define your categories
These actions and other editor actions are described in this section and in Populate Categories, on
page 73 and Manage Documents, on page 105.
Because multiple users can work on the same taxonomy, the Taxonomy module handles actions
performed to prevent any corruption of the taxonomy.
Edit a Taxonomy
Editing a taxonomy includes the following activities:
•
•
•
•
•
Add a Category
58
Delete a Category or Subtree
58
Restore a Category or Subtree
59
Rename a Category
59
Search a Category Node
59
HPE IDOL Collaborative Classifier (11.1)
Page 57 of 155
User Guide
Chapter 7: Author Taxonomies
•
•
•
Move a Category Subtree
60
Copy a Category Subtree
60
Copy a Resource Category
61
Add a Category
When you add a category, you add it as a child node to the current node. See Figure 3: Taxonomy
terminology, on page 46 to see the relationship between the current node and a child node.
To add category nodes to the taxonomy tree
1. In the taxonomy pane, select the category that you want to add the new child category to.
2. Click + to add the child category, or right-click and use the menu.
HPE ICC creates a locked category folder in the taxonomy.
3. Enter the name of the new child category, and then press Enter.
HPE ICC creates the new category node and shows another new category.
4. Continue adding nodes in the same way.
5. When you have added all the nodes that you want, press Escape or click anywhere in the
taxonomy pane except on another node.
HPE ICC creates and saves the new node or nodes.
NOTE:
To stop node creation at any time, click a node to discard your changes.
Name Categories
Category names identify the different categories in a taxonomy. The following rules apply to category
naming:
l
Categories at the same level in the taxonomy cannot have identical names.
l
Subcategories can use the same name as a parent.
Delete a Category or Subtree
When you delete a category or subtrees, they are marked as deleted. When you publish the parent
category, the category or subcategories are deleted.
To delete categories or subtrees in the taxonomy
1. In the taxonomy pane, select the category to remove.
2. Click - to mark a category for deletion.
A warning dialog box opens asking you to confirm the deletion.
3. Click Yes.
HPE IDOL Collaborative Classifier (11.1)
Page 58 of 155
User Guide
Chapter 7: Author Taxonomies
The node is shown with a line through the name.
NOTE:
To remove the subtree completely from view, you must publish the category.
Restore a Category or Subtree
If categories or subtrees are marked as deleted, they are shown in the taxonomy pane with a line
through the category name. If a category is still displayed with a line through the name, you can restore
it. After the parent category is published, the category or subtree is completely removed from the
taxonomy pane and cannot be restored.
To restore categories or subtrees that you deleted in the taxonomy
1. In the taxonomy pane, select the deleted category.
2. Right-click the category.
A shortcut menu opens.
3. Select Restore.
The node is restored and is shown without the line through the name.
Rename a Category
To preserve the subtree structure but change the name of the parent category, rename the category in
the taxonomy. You can also rename binary categories.
To rename categories in the taxonomy
1. In the taxonomy pane, select the category that you want to rename.
2. Make sure that the category is locked for editing. If it is not, click Edit
.
3. Right-click the category.
A shortcut menu opens.
4. Select Rename.
The text box becomes editable.
5. Enter the new name and press Enter.
The node is renamed.
For category naming rules, see Name Categories, on the previous page.
Search a Category Node
When you have large and complicated taxonomies, it might be faster to search for a node instead of
navigating to that node.
HPE IDOL Collaborative Classifier (11.1)
Page 59 of 155
User Guide
Chapter 7: Author Taxonomies
To search a category node
1. Select the category node you want to search.
2. Click Search node in tree
.
The Find Node In Tree dialog box opens.
3. Enter a search string in the Find field.
4. Use the Search text in and Search node contains lists to select your search preferences.
5. Click Find. The dialog box shows the Total Matches for the search string, and the first category
match in the node tree is highlighted.
6. To move through the series of matches, click Next or Previous. For each match, HPE ICC takes
you to the category node that matches the search string.
7. Click Close to end the search.
Move a Category Subtree
You can change the organization of your taxonomy manually by rearranging category subtrees . When
you move a category, the subtree for that category is moved with it.
To move a category subtree
1. Select the category.
2. Use a drag-and-drop operation to move the category with its subtree into the new location.
Copy a Category Subtree
You can copy a category subtree to create a new and separate subtree, or create a reference category.
See Reference Categories, on page 46 for more information about reference categories.
To copy a category subtree
1. In the categories pane, select the category to copy.
2. Right-click the category. Select Copy from the shortcut menu.
3. Select the location to add the copied category.
4. Right-click the category. Select Paste from the shortcut menu.
l
l
Select as Reference to create a reference category with the same name as the primary
category.
Select as Copy to copy the category as a new subtree.
5. The copied category is displayed in the taxonomy.
NOTE:
Reference category icons have small black arrows.
HPE IDOL Collaborative Classifier (11.1)
Page 60 of 155
User Guide
Chapter 7: Author Taxonomies
Reference Category Rules
You can modify only some of the fields for reference categories on the Category Details tab.
You can modify the following reference category properties independently of the primary categories.
l
The Description field on the Status secondary tab
l
The Reviewer Notes fields on the Status secondary tab
l
Any user-defined property fields on the User Defined secondary tab
l
The name of the reference category
NOTE:
When displaying category properties, the properties of the primary category are displayed for all
fields except in cases where the reference category has a value that overrides a primary
category value.
Some actions on a primary category result in different changes to the reference category.
l
Publishing a deleted primary category automatically removes any related reference categories.
l
Publishing a primary category does not automatically publish the reference categories.
l
Copying a subtree from a resource taxonomy maintains the reference if you also copy the primary
category.
If you copy the reference category without copying the related primary category, the reference
category becomes a standard category.
You can perform the following actions only on primary categories, not on reference categories:
l
Modify document lists
l
Modify training properties
l
Modify category rules
l
Recommend documents to the reference category
Copy a Resource Category
You can copy a category subtree from a resource taxonomy into an open taxonomy. Copying a
resource category allows you to use a category rule that you created for a category in another
taxonomy as a basis for the rule of a new category.
To open a resource category and copy a category subtree
1. In the Resources menu, click Open Resource Taxonomy.
The Open Resource Taxonomy dialog box opens.
2. Open the Select list.
The list of available resource taxonomies opens.
3. Select a resource taxonomy and click OK.
The taxonomy is displayed in the Resource Taxonomies pane.
HPE IDOL Collaborative Classifier (11.1)
Page 61 of 155
User Guide
Chapter 7: Author Taxonomies
NOTE:
You can have multiple resource taxonomies open. Access the open taxonomies from the list.
To review a resource taxonomy and copy a category subtree
1. Select the desired taxonomy.
The Resource Taxonomy and its category tree appear in the Resource Taxonomies pane.
2. To preview a category in the Resource Taxonomy, right-click a category.
The category working area pane changes to show the Category Rules and Category Details
tabs for the selected Resource Category as read-only.
3. Inspect the details on these two tabs to decide whether the category rule is useful or not.
4. If you want to use the resource category, use a drag-and-drop operation to move the category from
the Resource Taxonomy pane into the open taxonomy. The new category is displayed where you
placed it in the taxonomy.
NOTE:
You cannot copy the top-level root taxonomy in this way.
Edit Category Information
You can edit category node information on the Category Details tab. Use the three secondary tabs—
Status, User Defined, and IDOL Standard—for the following actions:
•
•
•
•
•
•
Add a Category Description
62
Change the Category Language
62
Add Reviewer Notes
63
View Reviewer Notes History
63
Create User-Defined Properties for a Category Node
64
Edit HPE IDOL Standard Properties
65
Add a Category Description
You can create a summary of the category and its contents.
To add or edit a category description
1. On the Status secondary tab, enter a description for the category into the Description text box.
2. Click Save Changes.
Change the Category Language
If your taxonomy includes categories that require different language encoding, you can change the
language on a category-by-category basis.
HPE IDOL Collaborative Classifier (11.1)
Page 62 of 155
User Guide
Chapter 7: Author Taxonomies
To change the category language
1. Open the Status secondary tab.
2. On the Status secondary tab, make sure that the category is locked for editing.
3. On the Category Language panel, click Change.
The Change Category Language dialog box opens.
4. The Language field displays the currently selected language. Choose the language you want to
use from the Select list.
5. Click OK.
The dialog box closes and the selected language is shown in the Language field on the Category
Language panel.
NOTE:
When you search for documents under Categorization, you can change the character
encoding for the category language. Character encoding is for search and training
purposes. See Search a Database, on page 78.
Add Reviewer Notes
Reviewers can leave notes for other reviewers and editors about the changes made during the
reviewing session.
To add reviewer notes
1. On the Status secondary tab, enter a note in the Reviewer Notes box.
2. Click Save Changes.
View Reviewer Notes History
You can inspect the history of reviewer notes for the category, and retrieve earlier notes. HPE ICC
shows the history of all saved reviewer notes that have been created by all users for the category.
The most recently created notes appear first in the list. If you want to find notes by a particular user,
click the User ID or Editor Comment column headers to sort the list alphabetically.
To view the history of reviewer notes and retrieve them
1. On the Status secondary tab, click Show History of Reviewer Notes
.
The History of Reviewer Notes dialog box opens.
2. Select the reviewer note that you want to use from the list.
The note is shown in the Selected Record pane.
3. Click Copy to Clipboard.
NOTE:
All the text in the review note is copied. Alternatively, you can highlight only part of the note
text in the Selected Record pane for copying.
HPE IDOL Collaborative Classifier (11.1)
Page 63 of 155
User Guide
Chapter 7: Author Taxonomies
4. Click Close.
5. In the Reviewer Notes box, press Ctrl-V to paste the reviewer note.
Create User-Defined Properties for a Category Node
Add a User-Defined Property
When you create a user-defined property, you can add it to a single category, or to the entire
subtree.When you apply the user-defined property to the subtree, the property is applied to all the
children of that category.
To add a user-defined property
1. On the User Defined secondary tab, click +.
2. Enter the property name and value in the Add Property dialog box.
3. (Optional) Click Apply to the entire subtree.
4. Click OK.
The new property is displayed in the User Defined property list.
Edit a Property
You can edit all the properties of an existing property. You can also apply the changed property to all the
children in the subtree.
To edit a property
1. On the User Defined secondary tab, select a property in the User Defined property list.
2. Click Edit.
3. Edit the fields in the Add Property dialog box.
4. (Optional) Click Apply to the entire subtree to apply the property to all child nodes.
5. Click OK.
Delete a User-Defined Property
If you delete a property from the parent category, you can delete it from the parent category only, or
delete it from the entire subtree.
If you delete a property from one of the child categories, HPE ICC deletes the property from that one
category. When you delete a property, it is automatically removed from all categories that it was
applied to.
To delete a user-defined property
1. On the User Defined secondary tab, select the property
2. Click -.
3. If you are editing a parent category, a confirmation dialog box opens. Click either Yes or No to
HPE IDOL Collaborative Classifier (11.1)
Page 64 of 155
User Guide
Chapter 7: Author Taxonomies
apply the change to the entire subtree.
l
Yes. The property is removed from all categories in the subtree.
l
No. The property is removed from the parent category only.
Edit HPE IDOL Standard Properties
You can edit system-generated properties on the IDOL Standard secondary tab. These properties
affect category queries.
Edit Number of Results Returned by a Query
NOTE:
This procedure does not control the number of returned results that are displayed by default in
HPE ICC.
You can control the maximum results displayed on each page by setting the Size value to the
left of the page navigation in the results section of a search page.
For the Document tab, you can control how many results are returned by setting the
category.query.max.result property in the acc.properties file. For more information, see
Enable Paging Managers, on page 20.
To edit the number of results
1. On the IDOL Standard secondary tab, select NumResults in the IDOL Standard property list.
2. Click Edit.
3. Edit the value in the Add Property dialog box for the number of results you want for a category
query. The default value is 6.
4. Click OK.
Edit Databases
Use this procedure to edit databases that your HPE Category server will pull documents from.
To edit the databases for a category query
1. On the IDOL Standard secondary tab, select Databases in the IDOL Standard property list.
2. Click Edit.
3. Edit the value in the Add Property dialog box for the databases you want to use for category
queries. A default value is not defined, which means that a category query applies to all databases
in HPE IDOL Server.
4. Click OK.
HPE IDOL Collaborative Classifier (11.1)
Page 65 of 155
User Guide
Chapter 7: Author Taxonomies
Add and Maintain Users
You can add users to each category, or to entire subtrees, from the Permissions tab. You can add a
user as a publisher, editor, or reviewer, and edit those user roles later.
Add a User
HPE IDOL Collaborative Classifier users are assigned roles and, based on their roles, different
permissions. For more information about roles and permissions, see Manage User Permissions, on
page 36.
NOTE:
To add users to binary categories, you must first create the users through the Administration
interface. See Add and Maintain Users, above.
To add users to categories and entire subtrees
1. Click +.
The Choose User dialog box opens.
2. Enter the ID of the new user in the User or Domain\User field.
3. Select the user role from the Role list.
4. Select Apply to the entire subtree if you want to add the user to all subcategories.
NOTE:
This selection is available only if the category has subcategories.
5. Click OK.
The user ID and role appear in the Users list.
Edit User Privileges
You can only edit the role for users with the same or fewer privileges than your role. You cannot assign
someone a privilege level greater than the one that you hold. For example, if you are an editor, you
cannot change the privileges of a publisher on that category node. You can change the privileges for
reviewers and other editors, but you cannot promote those users to the publisher role.
NOTE:
An administrator is not restricted to the user roles. An administrator can perform any action in
HPE ICC.
To edit user privileges
1. Select a user from the Users list.
2. Click Edit.
3. Change the role for that user in the Role list.
HPE IDOL Collaborative Classifier (11.1)
Page 66 of 155
User Guide
Chapter 7: Author Taxonomies
4. Check Apply to the entire subtree if you want to change the user privileges to all subcategories.
5. Click OK.
Remove a User
If you have the same or greater user privileges, you can remove a user from a category.
To remove a user
1. Select a user from the Users list.
2. Click -.
A confirmation dialog box opens.
3. Select Remove from the entire subtree if you want to remove the user from all subcategories.
4. Click Yes.
About Filters
You can set filters to view only those categories that meet certain requirements. For example, you can
set the Locked filter to view only those categories that are locked.
You can also create custom views to further narrow the information that you want to view. For example,
you can create a custom view to see only those categories for which you have publisher rights.
NOTE:
Filters require an explicitly-defined user role. If administrators need to use filters, add them to
the taxonomy as publishers.
To use the filters already available in HPE IDOL Collaborative Classifier, select a filter from the Filters
menu. When you select a filter, HPE ICC switches to the Filtered Draft View and the taxonomy pane
displays the categories that fit the filter requirements.
For the available filters and their descriptions, see Taxonomy Filters Menu, on page 28.
Create Custom Filters
You can use custom filters in HPE ICC to create individual views of information in a taxonomy. You
can use filters to determine specific tasks that you want to perform on a regular basis. For example,
you might want to view only the category nodes in which you have editor privileges.
To create a custom filter
1. In the Filters menu, select Create Custom.
The Create Custom Filter dialog box opens.
2. Make your selections. You can include:
l
Categories that you or another user have locked
l
Categories that you or another user can review, edit, or publish
HPE IDOL Collaborative Classifier (11.1)
Page 67 of 155
User Guide
Chapter 7: Author Taxonomies
l
Categories with or without documents to be reviewed
l
Modified categories
3. (Optional) Enter a name for the filter.
NOTE:
If you create a custom filter with the same name as an existing filter, the new filter
overwrites the old filter.
4. Click OK.
Remove Custom Filters
You can delete custom filters that you no longer need. When you delete a filter, it is completely
removed from the system and you cannot restore it.
To remove all custom filters
1. In the Filters menu, select Remove Custom.
A confirmation dialog box opens.
2. Click Yes to remove all custom filters.
View Published Categories
If you want to view only the categories that have been published, you can change the view.
To view the published state of categories in a taxonomy
1. Open a taxonomy.
The open taxonomy is displayed in the draft view.
2. In the View menu, select Published View.
l
The taxonomy pane now displays only the published categories in that taxonomy.
l
The document pane now displays only the Documents and Category Details tabs.
3. Select one of the published categories.
The documents list displays the published documents.
HPE IDOL Collaborative Classifier (11.1)
Page 68 of 155
User Guide
Chapter 8: Author Binary Categories
Chapter 8: Author Binary Categories
This section discusses the binary category authoring tasks performed by users who have editor
privileges. Most of the actions that you can perform on binary taxonomies are the same for regular
categories, with a few exceptions. These differences are outlined in this section.
•
•
•
•
About Binary Taxonomy Authoring
69
Edit Binary Categories
69
Edit Binary Category Information
70
Manage Users
71
About Binary Taxonomy Authoring
Unlike regular taxonomies, there is only one binary taxonomy. When enabled, it is displayed in the
Summary window with the title Binary Taxonomy.
You cannot create additional binary taxonomies. You must use the single binary taxonomy to create
and manage multiple binary categories. Unlike regular taxonomies, individual binary categories cannot
have child categories; all binary categories are listed as first-level children of the Binary Taxonomy
root.
Edit Binary Categories
You can perform the following actions on a binary category:
l
Add a Binary Category
l
Delete a Binary Category from the Taxonomy
l
Rename a Category
Add a Binary Category
You can add binary categories to the Binary Taxonomy root as first-level children.
HPE IDOL Collaborative Classifier (11.1)
Page 69 of 155
User Guide
Chapter 8: Author Binary Categories
To add a binary category
1. In the taxonomy pane, select Binary Taxonomy.
2. Click +.
A locked category icon is displayed in the taxonomy.
3. Enter the name of the new binary category and press Enter.
The new category node is created and another new category is opened.
4. When you have added the nodes that you want, press Escape or click anywhere in the taxonomy
pane but not on another node.
HPE ICC creates and saves the new node or nodes.
Delete a Binary Category from the Taxonomy
To delete a binary category
1. In the taxonomy pane, select the category that you want to remove.
2. Click -.
A confirmation dialog box opens.
NOTE:
If the category is already published on the host server, and you have administrator
privileges, a checkbox allows you to delete the category from HPE IDOL.
3. Click Yes.
The binary category is removed from the taxonomy.
Edit Binary Category Information
You can edit category node information on the Category Details tab. Use the two secondary tabs—
Status and IDOL Standard—for the following actions:
l
l
l
l
Add a Binary Category Description. See Add a Category Description, on page 62 for more
information.
Change the Binary Category Language. See Change the Category Language, on page 62 for more
information.
Add or view Reviewer Notes. See Add Reviewer Notes, on page 63 and View Reviewer Notes
History, on page 63 for more information.
Edit system-generated properties. See Edit HPE IDOL Standard Properties, on page 65 for more
information.
HPE IDOL Collaborative Classifier (11.1)
Page 70 of 155
User Guide
Chapter 8: Author Binary Categories
Edit HPE IDOL Standard Properties
You can edit system-generated properties on the IDOL Standard secondary tab. The first two
properties, Databases and NumResults, affect category queries, just as in regular categories. The
other options affect only binary queries (see Binary Queries, on page 103).
You can modify the following properties:
Databases
The databases used for queries.
NumResults
The number of results returned by queries.
BiasFactor
The degree of bias in favor of negative or positive results.
DefaultTermWeight The default weight assigned to an unrecognized test term.
LowerWeightLimit
The minimum weighting that a term can receive.
MinDocOccs
The minimum number of documents in which a term must appear before it can
be weighted and used to score.
MinTestTerms
The minimum number of (scoring) terms a document must contain before a
positive decision is allowed.
TestTermsPerDoc
The maximum number of 'best' terms to use from each test document for
weighting.
TestThreshold
The percentage that must be scored before a document is flagged as positive.
UpperWeightLimit
The maximum weighting that a term can receive.
For more detailed explanations of these properties, see the HPE IDOL Server Reference.
To edit an HPE IDOL property
1. In the IDOL Standard property list, select the property that you want to modify.
2. Click Edit.
An Edit Property dialog box opens.
3. Add or change the property value in the Value field.
4. Click OK.
The new value is displayed in the property list.
Manage Users
You can add users to each category, or to entire subtrees, from the Permissions tab.You can add a
user as a publisher, editor, or reviewer, and edit those user roles later.
You can perform the following actions for Binary taxonomy users:
l
Add a user. See Add a User, on page 66 for more information.
l
Edit user privileges. See Edit User Privileges, on page 66 for more information.
HPE IDOL Collaborative Classifier (11.1)
Page 71 of 155
User Guide
Chapter 8: Author Binary Categories
l
Remove a user. See Edit User Privileges, on page 66 for more information.
HPE IDOL Collaborative Classifier (11.1)
Page 72 of 155
User Guide
Chapter 9: Populate Categories
Chapter 9: Populate Categories
This section describes how to populate the categories that define your taxonomy.
•
•
•
•
•
•
•
•
Introduction
73
Select a Categorization Method
73
Search a Database
78
Create a Subtaxonomy
79
Define Category Rules by Using Training Documents
81
Define Category Rules by Using a Query
85
Manually Assign Documents
97
About Automatic Query Guidance
98
Introduction
The first step towards populating categories is to use the methods provided on the Categorization tab
to build a category. The Training and Query methods allow you to define a category rule for
classification of documents in a subject matter domain. The category rule determines which
documents in a collection are included in the category. The Manual method allows you to add to the
category by assigning extra documents after the rule has populated the category.
NOTE:
HPE recommends that only people with some subject-area expertise categorize information.
The next step is to test the category rule against a benchmark by using the QA tab. You can evaluate
how the rule performs when used to search for documents and populate the category.
After you have populated the category, you can view the documents in the Documents tab, where you
can finalize the set of documents to publish under the category. See Manipulate Documents in the List,
on page 111.
Select a Categorization Method
To choose a Categorization method, select one of the Select Categorization Method options on the
Categorization tab:
l
l
Training. Search a collection and select documents from the result set to use to train your category
for a specific category rule. Use training documents, text, or terms and weights to improve your
training.
Query. Compose a query expression that forms a category rule.
Use the Query Text and Filter Text wizards to build your query expression. You can also create your
query from rules of child categories.
l
Manual. Search a collection and manually assign documents from the search result set to a
category or categories.
HPE IDOL Collaborative Classifier (11.1)
Page 73 of 155
User Guide
Chapter 9: Populate Categories
The layout of the Categorization tab is based on workflows for the method you choose. Each method
is described in the following sections.
You can use HPE IDOL’s Automatic Query Guidance (AQG) functionality with the Training or Query
methods if it is enabled. See About Automatic Query Guidance, on page 98.
Training
Use the Training method to train a category for a category rule. The objective is to define a category
rule that determines which documents populate the category.
You can also conduct training based on
l
An example set of documents combined with a piece of training text.
l
A piece of training text alone, which becomes a single training document.
You can further refine the list of terms and weights that define the category rule by adding or removing
terms from the list, or by editing the assigned weight.
HPE IDOL Collaborative Classifier (11.1)
Page 74 of 155
User Guide
Chapter 9: Populate Categories
To train your category to define a category rule
1. In the Taxonomy pane, select the category that you want to train. On the Categorization tab,
select the Training method.
2. In the Search area of the Training working pane, search one or more databases. See Search a
Database, on page 78.
3. Examine the search result documents in the upper half of the pane.
4. Choose the documents that you think are good examples of the subject matter that you want your
category to contain.
5. Select the preferred documents for training.
6. Enter a piece of training text if required.
7. Train the category. See Define Category Rules by Using Training Documents, on page 81.
8. (Optional) Create a subtaxonomy to further refine your category. See Create a Subtaxonomy, on
page 79 for more information.
9. Refine the resulting category rule if required.
Query
Use the Query method on the Categorization tab to define a category rule by building a query text
expression. The category rule determines which documents populate the category from any specified
collection of documents.
You can build a query text in several ways:
l
Enter a query text directly into the query working area.
l
Use the Query Text and Filter Text wizards.
l
Review a History of Query Texts and Filters and retrieve any saved query or filter expression.
l
Use the rules from child categories.
l
Use Automatic Query Guidance (see About Automatic Query Guidance, on page 98).
HPE IDOL Collaborative Classifier (11.1)
Page 75 of 155
User Guide
Chapter 9: Populate Categories
To use a query and define a category rule
1. Use the Query tab to build a query. See Build a Query Text, on page 85.
TIP:
You can start building a query or creating a filter by retrieving a saved query or filter. See
About Query Text and Filter History, on page 95.
You can also import a query rule from a child category. See Rules for Child Categories, on
page 97.
2. Use the Filter tab to add any required filters to the query. See Build a Query Filter, on page 93.
3. Test the query. See Test a Query, on page 95.
4. Inspect the set of documents returned by the query.
5. Refine the query if required.
6. Repeat Step 2 to Step 5 until you are satisfied with the query.
7. Save the query to define the category rule.
8. (Optional) Create a subtaxonomy to refine your category further. See Create a Subtaxonomy, on
page 79 for more information.
9. (Optional) You can review documents generated by the category rule on the Documents tab and
finalize the documents that the category contains when the category is published.
HPE IDOL Collaborative Classifier (11.1)
Page 76 of 155
User Guide
Chapter 9: Populate Categories
Manual
Use the Manual method to assign documents to a category manually. You must determine whether a
document’s subject matter qualifies it for inclusion in the category, rather than rely on automatic
selection performed according to a category rule. HPE therefore recommends that you use the Manual
method only for categories in which you have expert knowledge of the subject matter.
If a category rule already exists for the category, the Documents tab generates a list of documents that
comply with the category rule. You can then use the Documents tab to finalize which documents are
published under the category. See Manage Documents, on page 105.
To assign documents to a category manually
1. On the Categorization tab, select Manual. Use the upper half of the pane to perform a search.
See Search a Database, on the next page.
2. When you have an adequate set of documents in the search area, you can select documents from
the search results and add them to the Documents tab.
NOTE:
If the category already has a category rule defined when you conduct a search, the
documents in the search result set that comply with the rule automatically populate the
Documents tab.
HPE IDOL Collaborative Classifier (11.1)
Page 77 of 155
User Guide
Chapter 9: Populate Categories
3. Choose one of the three options:
l
Assign Documents to Accepted
l
Assign Documents for Review
l
Assign Documents to Excluded
TIP:
To assist selection, you can view document details in two ways:
o
Click Show/Hide summary of Documents
.
o
Double-click a document to launch a browser window with a three-pane view.
You must configure viewing to use this function. See Configure the Taxonomy Module,
on page 18.
4. (Optional) After you assign documents, you can use the Manual Action Results pane to change
the assignments. See About the Manual Action Results Pane, on page 98.
5. (Optional) To help focus on suitable documents in a results list, you can select documents and
create a subtaxonomy from them. See Create a Subtaxonomy, on the next page.
See Manually Assign Documents, on page 97 for more information on manually assigning documents.
Search a Database
You can search the documents in a database from the Categorization tab.
The search area occupies the upper half of the working area pane for the Manual and Training
categorization methods, and the lower half for the Query method. The procedure for searching a
database is similar for each method.
TIP:
You can set the maximum limit of documents retrieved by a search by changing the default
value of the following parameter in the acc.properties file:
search.max.doc.number=10000
HPE ICC uses this number to validate the number typed by a user in the Show maximum of
text field above a Search panel. If a user enters a number greater than the above limit and clicks
Search, an error message appears, and the user must enter a valid number.
To search a database
1. Click the Categorization tab.
2. Select a categorization method.
3. In the search area, click Database.
The Database menu opens.
4. Select the databases that you want to search, or select All Databases.
5. Enter values into the search specification fields.
HPE IDOL Collaborative Classifier (11.1)
Page 78 of 155
User Guide
Chapter 9: Populate Categories
l
Encode. Specifies the character encoding.
NOTE:
The available options displayed in the Encode list are dependent on the language set for
the category. See Change the Category Language, on page 62.
l
Show maximum of. This option specifies the maximum number of documents to retrieve.
l
With a score above. This option specifies the score above which documents are retrieved.
6. If you are using the Training or Manual method, enter the expression that you want to search for
in the Search or Research text box. If you are using the Query method, enter the expression in
the Query working pane.
7. If you are using the Training or Manual method, click Search.
8. If you are using the Query method, click Test Query
.
Any documents that match the query search criteria appear in the document list.
Search by Saved Searches
If you are using the Training or Manual method, you can also search by using saved searches.
To search by using a saved search
1. Configure your search as described in the previous section.
2. Instead of entering a search word, click the arrow at the right margin of the Search text box.
3. Select one of the Saved Searches from the list.
4. Click Search.
Any documents matching the search criteria appear in the document list.
Create a Subtaxonomy
When you search a database by using the Training or Manual method, you can create a subtaxonomy
by using:
l
A set of documents selected from the search results
l
The results of a specified query
A subtaxonomy is generated automatically, based on the documents that you select or those returned
by a search. The subtaxonomy becomes a subtree in the selected category. This hierarchical category
subtree provides a preview of the main concepts found in the set of documents used to create the
subtaxonomy.
Use the subtaxonomy as a heuristic tool that helps you to determine a category rule that is required to
define a category. For example, if you are working in a subject area about which you have limited
knowledge, the subtaxonomy helps you focus on the area of crucial concern in the wider domain
covered by the set of documents. You can manually eliminate specific nodes from the subtaxonomy
HPE IDOL Collaborative Classifier (11.1)
Page 79 of 155
User Guide
Chapter 9: Populate Categories
tree that lie outside the main area of concern for your category, leaving core concepts that can then
inform the category rule that you are defining.
Create a Subtaxonomy from Selected Documents
To create a subtaxonomy
1. On the Categorization tab, select the Training or Manual categorization method.
2. Search one or more databases.
3. In the search result set, select the documents that you want to include in the subtaxonomy.
4. Click Create Subtaxonomy
.
The Subtaxonomy Wizard opens.
5. Select From Selected Documents.
6. Click Finish.
A Confirm Actions dialog box opens.
7. Review the setting and click Create Subtaxonomy, or click Exit to exit the Subtaxonomy
Wizard without creating the new subtaxonomy.
8. Click Refresh Category after your subtaxonomy is created.
HPE ICC creates the new subtaxonomy as a child of the currently selected category.
Create a Subtaxonomy from Search Results
To create a subtaxonomy from search results
1. On the Categorization tab, select the Training or Manual categorization method.
2. Select one or more databases.
3. Click Create Subtaxonomy
.
The Subtaxonomy Wizard opens.
4. Select From Search Results.
5. Click Next.
The Set Search Parameters dialog box opens.
6. In the Query field, enter the query that you want to use to generate results for the subtaxonomy.
Click Finish.
The Confirm Actions dialog box opens.
7. Review the subtaxonomy and click Create Subtaxonomy.
NOTE:
The specified query is performed independently of any search that you have already
conducted in the search area.
8. Select Refresh Category after your subtaxonomy is created.
HPE IDOL Collaborative Classifier (11.1)
Page 80 of 155
User Guide
Chapter 9: Populate Categories
HPE ICC shows the new subtaxonomy as a child of the currently selected category.
NOTE:
Auto-generation assigns a name to the subtaxonomy. You can rename the subtaxonomy.
Define Category Rules by Using Training Documents
With this method, you compile a set of example documents, which are chosen on the basis of their
subject matter content. When you train the category by using the example documents, HPE ICC
defines a category rule for the category. The rule consists of an analysis of the example documents
that produces a list of terms with weights assigned to each term. The category rule determines which
documents populate the category from any specified collection of documents.
Mark Documents for Training
You must mark certain documents as training documents. Use documents that are most representative
of that category of the taxonomy.
To mark documents for training
1. Examine the documents returned in the search result set. To see more details of these documents
to assist your selection:
l
Click Show/Hide summary of Documents
.
l
Double-click a document to launch a browser window with a three-pane view.
NOTE:
You must configure viewing to use this function. See Configure the Taxonomy Module, on
page 18.
2. Select the documents that you want to use as relevant training documents.
3. Click Relevant for Training
.
The documents appear in the lower half of the Training working pane under the Training
Documents tab.
NOTE:
You can search again and repeat this process with a new search result set until you obtain
an adequate set of training documents.
HPE IDOL Collaborative Classifier (11.1)
Page 81 of 155
User Guide
Chapter 9: Populate Categories
Remove Documents from the Training Documents List
To remove documents from the training documents list
1. Select the documents that you want to remove from the training list.
2. Click Remove from Training
.
Train the Category
To train your category, HPE recommends that you have at least 20 relevant documents in the training
area to define a meaningful category rule. On the Training Text tab, you can also enter a piece of text
as an additional example for training.
You can train your category based on:
l
l
The example set of documents alone
Training text alone, where the piece of training text can be a single word, a string of text, or a larger
piece of text copied from a separate document
l
A combination of example documents and training text
l
The results of Automatic Query Guidance (see About Automatic Query Guidance, on page 98)
To train your category
1. If you are using example documents, assemble a set of relevant documents in the training area.
2. If you are using training text, click the Training Text tab.
3. Enter the text to use as training text. Click Save.
4. Click Train.
HPE ICC trains the category for the example documents and training text.
When the process is complete, the Terms and Weights tab opens.
Refine a Category Rule by Using Terms and Weights
On the Terms and Weights tab, the category rule is shown as a list of terms with weights in
descending order. The numerical values assigned as weights to each term are a measure of relevance.
Terms and weights generated automatically by HPE IDOL and user-created terms are treated
differently. They appear in different colors in the term list, and the procedure for deleting each type of
term is different.
You can refine the category rule in several ways:
l
Add a new term
l
Remove a term
l
Edit a term and its weight
HPE IDOL Collaborative Classifier (11.1)
Page 82 of 155
User Guide
Chapter 9: Populate Categories
l
Keep a term
l
Add user-defined properties that control how the terms and weights are generated
Add or Edit a Term
To add a new term
1. Click Add Terms and Weights (+).
The Add Term dialog box opens.
2. Enter the term to add, and a corresponding weight.
3. Click OK.
HPE ICC adds the new term to the list at the appropriate point in the weight list. New terms appear
in bold blue italicized text, and the row is grayed out.
To edit a term
1. In the Terms and Weights list, select the term to edit.
2. Click Edit. If the selected term is generated by HPE IDOL, a warning message opens. Click Yes
to continue.
The Edit Term dialog box opens.
3. Edit the term or the weight value.
4. Click OK.
HPE ICC shows the edited term in the list at the appropriate point in the weight value order, and
the row is grayed out. User-generated terms appear in bold blue italicized text; terms generated by
HPE IDOL appear in bold red italicized text.
NOTE:
You can remove a selected user-generated term in the normal way. To remove a term
generated by HPE IDOL, see To remove a term generated by HPE IDOL, on the next
page.
Keep a Term
When you publish a category, all the terms are regenerated for the set of training documents that you
have assembled, including any terms that you removed when you edited the Terms and Weights tab.
Use the Keep button to retain the terms that you want to keep when you publish the category. The
terms that you removed are lost by removing the training documents after you have edited and before
you publish the category.
To keep a term
1. Select the term that you want to keep for the published category in the Terms and Weights tab.
2. Click Keep.
HPE ICC shows the selected term in bold italicized text, and the background of the row changes
to gray. User-generated terms appear in blue text; terms generated by HPE IDOL appear in green
text.
HPE IDOL Collaborative Classifier (11.1)
Page 83 of 155
User Guide
Chapter 9: Populate Categories
You can remove a selected user-generated term in the normal way. To remove a term generated
by HPE IDOL, see To remove a term generated by HPE IDOL, below.
Remove Terms
If a term does not fit the category, you can remove it to return better document results.
NOTE:
If you remove a term generated by HPE IDOL from the list, its weight is set to 0.
If you want to completely remove a term generated by HPE IDOL from the list, you must first
retrain the category without documents. See To remove modified terms, below.
To remove a user-generated term
1. In the Terms and Weights list, select the term to remove.
2. Click Remove Terms and Weights (–).
3. In the confirmation dialog box, click Yes.
HPE ICC removes the selected term from the list.
To remove a term generated by HPE IDOL
1. In the Terms and Weights list, select the term to remove.
2. Click Remove Terms and Weights (–).
3. In the confirmation dialog box, click Yes.
HPE ICC shows the term at the bottom of the term list with a weight of 0, in bold red italicized text.
The row is grayed out.
To remove modified terms
1. Remove all training documents. See To remove documents from the training documents list, on
page 82.
2. Click Train.
The only terms that remain in the Terms and Weights list are terms that have been marked
“Keep,” user-generated terms, terms added from document best terms (see About Document Best
Terms, on the next page), and terms generated by HPE IDOL that have been modified or removed.
3. Select the terms that you want to remove.
4. Click Remove Terms and Weights (–).
HPE ICC removes the selected terms from the list.
User-Defined Term Properties
You can add up to four user-defined properties to the category, which enable you to fine-tune how the
terms and weights are generated. Add these properties to the category by using the User Defined
secondary tab on the Category Details tab. See Add a User-Defined Property, on page 64.
l
MaxTerms. Set the maximum number of terms for the category. The default is 100, and the
permitted range of values is 1 to 4096.
HPE IDOL Collaborative Classifier (11.1)
Page 84 of 155
User Guide
Chapter 9: Populate Categories
l
l
l
MinTermLength. Set the minimum length of a term to include in the category. The default is 0, and
the permitted range of values is 1 to 4096.
MinTermWeight. Set the minimum weight of a term to include in the category. The default is 0, and
the permitted range of values is >1.
Attenuation. When added to the category, the weights of terms that are less important to the
category are reduced. The higher the setting for attenuation, the greater the reduction of a term’s
weight as its importance decreases. The permitted range of values is 0 to 20. A setting of 0 reverts
to an internal default.
About Document Best Terms
You can view and retrieve best terms from any document from the search results or the document list.
The best terms are the conceptually most important terms that the document contains.
To view document best terms
1. Select a document from the search results or the document list.
2. Right-click, and then click Show Best Terms.
The View and Copy Selected Document Best Terms dialog box opens.
3. Select the rows that contain the terms that you want to use.
4. Click Copy Selected Terms.
The selected term is copied to the clipboard.
5. Click Close.
After you copy one or more terms, you can:
l
Go to the Terms and Weights tab and paste a single term directly into the Add Term dialog box.
l
Go to the Query tab and paste the set directly into the work pane to form the basis of a query.
l
Paste the set into a text editor to sort before using.
Define Category Rules by Using a Query
Use the Query method on the Categorization tab to define a category rule by building a query text
expression. The category rule determines which documents populate the category from any given
collection of documents.
Build a Query Text
On the Categorization tab, select the Query method.
Enter Query Text Directly
If you are familiar with HPE IDOL queries, you can enter query text directly.
HPE IDOL Collaborative Classifier (11.1)
Page 85 of 155
User Guide
Chapter 9: Populate Categories
To enter a query directly
1. Click the Query tab.
2. Click in the working area and enter your query text.
The Save Query button flashes, prompting you to save the modified text.
3. Open the Filter tab and add any required filter fields to the query.
The Save Filter button flashes, prompting you to save the modified text.
4. Return to the Query tab.
Use the Query Text Wizard
You can enter a basic query directly into the Query tab, and then use the wizard to build and manage a
further query and append it to your original to form a complex query.
For each type of query text, there is a specific Create Statement panel on the wizard.
To use the Query Text wizard to build a query
1. Select the text type.
l
l
l
l
Field Restriction. Place a restriction on the query with respect to a field. See Build a Query by
Using a Field Restriction Query Type, below.
Fuzzy. Enter a string that is close in spelling to the word that you want to query.
Soundex Keyword. Enter a word that is not the correct spelling of the word you want to query,
but is phonetically similar to it. See Build a Query by Using a Soundex Keyword Query Type,
on page 89.
Use Existing Rule. Use a query text that is the category rule defined for another category.
See Build a Query by Using an Existing Rule Query Type, on page 90.
2. Use the wizard to create a query for that text type.
3. Append the query. See Append a Query, on page 91.
Build a Query by Using a Field Restriction Query Type
When you build a query by using the field restriction text type, the query returns only documents that
contain a specific value in a specific field.
You can also use the wizard to create a field restriction query and append it as a subquery to a main
query text. If you do this, those documents returned for the main query text and which satisfy the field
restriction subquery are assigned a higher relevance.
To build a query by using a field restriction query type
1. Click Open Query Text Wizard.
The Manage Query Text dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 86 of 155
User Guide
Chapter 9: Populate Categories
2. From the Query Text Type list, select Field Restriction.
3. Inspect the Format and Examples fields.
l
l
l
The Format field gives the correct formatting for a query with field restriction for use by
HPE IDOL.
The Examples field gives an example using correct formatting.
For further information on field restriction queries, see the HPE IDOL Server Administration
Guide.
4. Enter some query text into the Text field.
5. In the Set Fields panel, select a field to which you want to restrict the query from the Select Field
list.
6. Click Add.
The field is shown in the Selected text box.
NOTE:
You can add more than one field restriction to a query text.
7. Click Create.
HPE ICC creates the field restriction query.
HPE IDOL Collaborative Classifier (11.1)
Page 87 of 155
User Guide
Chapter 9: Populate Categories
Build a Query by using a Fuzzy Query Type
A fuzzy query returns documents that contain words that are close in spelling to the string entered for
the fuzzy field.
You can also use the wizard to create a fuzzy query to combine with a query text. Alternatively, you
can use the wizard to create a simple fuzzy query, and combine this query with a main query text on the
Query tab.
To build a query by using a fuzzy query type
1. Click the Open Query Text Wizard icon.
The Manage Query Text dialog box opens.
2. From the Query Text Type list, select Fuzzy.
3. Inspect the Format and Examples fields.
l
The Format field gives the correct formatting for a fuzzy query for use by HPE IDOL.
l
The Examples field gives an example of correct formatting.
For further information on fuzzy queries see HPE IDOL Server Administration Guide.
4. Enter some query text into the Text field if required.
5. In the DREFUZZY field, enter a fuzzy query term.
For example, if you enter “best selling author” for the query Text field and “Rowlling” in the
DREFUZZY field, your query might return for documents containing the text such as:
l
best selling author Rowling.
l
best selling author Rowllin.
HPE IDOL Collaborative Classifier (11.1)
Page 88 of 155
User Guide
Chapter 9: Populate Categories
l
best selling author Rowlline.
What is returned for a fuzzy query also depends upon the fuzzy level tolerance set for the
query.
6. In the Fuzzy Level field, enter the spelling tolerance level.
The tolerance level is a measure of closeness in spelling to the specified string in the DREFUZZY
field. A low fuzzy tolerance level setting, for example 1, returns only words that are very close in
spelling. A high setting, for example 6, returns words only slightly similar in spelling.
NOTE:
HPE recommends that you do not set a fuzzy tolerance level greater than 6. HPE ICC
could return results that are unrelated to the query if the fuzzy matching is too flexible.
DREFUZZY internally determines a default spelling tolerance level when it judges whether words
in a result document are similar enough to the words that are specified in the query. If you do not
use the default, set a level from 1 to 7.
7. Click Create.
HPE ICC creates the fuzzy query.
Build a Query by Using a Soundex Keyword Query Type
A soundex keyword query returns documents that contain the keyword entered, and documents that
contain words phonetically similar to the keyword.
You can create a soundex query that contains one or more keywords.
You can also use the wizard to create several successive soundex queries, each time appending the
soundex query to the main query. This process builds a main query expression on the Query tab
containing several soundex terms.
To build a query by using a soundex query type
1. Click Open Query Text Wizard.
The Manage Query Text dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 89 of 155
User Guide
Chapter 9: Populate Categories
2. From the Query Text Type list, select Soundex Keyword.
3. Inspect the Format and Examples fields.
l
l
The Format field gives the correct formatting for a soundex keyword query for use by
HPE IDOL.
The Examples field gives an example of correct formatting.
For further information on soundex keyword queries, see the HPE IDOL Server Administration
Guide.
4. In the Soundex field, enter a keyword query term. Click Create.
HPE ICC creates the soundex query text.
Build a Query by Using an Existing Rule Query Type
You can import the query text that defines the category rule for another category and use it to build a
query.
To create a complex query, repeat the import and use Boolean operators to join multiple queries used
by other categories.
To build a query by using an existing rule query type
1. Click Open Query Text Wizard.
The Manage Query Text dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 90 of 155
User Guide
Chapter 9: Populate Categories
2. From the Query Text Type list, select Existing Rule.
NOTE:
The Format and Examples fields are not used for this option.
3. In the Find Category field, enter the name of the category whose rule you want to import.
4. Click Go.
HPE ICC populates the Select list with categories with the specified name and which have a
category rule defined.
5. Select the category whose rule you want to import.
The selected category rule is shown in the Selected field.
6. Click Create.
HPE ICC creates the query that defines the category rule for the selected category.
Append a Query
Appending a query is a two-stage process:
l
l
First stage. The query is appended in the wizard and moved initially into a subquery area. This
process allows you to create another query in the wizard, using either the same text type or another
text type. You can then append the second query to the first to form a single complex subquery.
Second stage. The subquery is appended to the main query. You can use the wizard repeatedly and
append successive subquery parts to assemble a single but complex main query.
HPE IDOL Collaborative Classifier (11.1)
Page 91 of 155
User Guide
Chapter 9: Populate Categories
The procedure for appending a query is the same for each query text type.
To append a query
1. Build a subquery so that it is shown in the Create field.
2. In the Append Statement to Sub Query with... panel, click Append.
HPE ICC moves the query to the subquery area.
3. To build a complex subquery, leave the first query in the subquery area in the wizard, select the
text type and build another query. If you want to append the query to the main query immediately,
go to Step 6.
4. When the second query is shown in the Create field, select the Boolean operator to join the
second query to the first query in the subquery area. Click Append.
HPE ICC appends the second query to the first to form a complex subquery.
5. Repeat Steps and 3 and 4 until the complex subquery is sufficient.
6. In the Append Sub Query to Query Text with... panel, select a Boolean operator to join the
subquery to the main query text. At the bottom of the main wizard, click Append.
If there is already query text, the subquery is appended to it. If there is no query text, the subquery
HPE IDOL Collaborative Classifier (11.1)
Page 92 of 155
User Guide
Chapter 9: Populate Categories
is transferred into the main query area.
The Save Query button flashes, prompting you to save the modified text.
For more information on Field Restriction, Fuzzy and Soundex Keyword for query statements, see the
HPE IDOL Server Administrator Guide.
Build a Query Filter
Use the Filter tab to restrict the set of documents returned by the query. You can use the FieldText
parameter to restrict the documents that a query returns to those that contain a specified value in a
specified field.
For more information on the FieldText parameter in query statements, see the HPE IDOL Server
Administration Guide.
You can use the FieldText wizard on the Filter tab to create FieldText query filter statements. You can
build complex query filter statements from several different field specifiers and, by using the append
function, concatenate the filter expressions.
To create a FieldText query filter
1. Click Open FieldText Wizard
.
The Manage FieldText dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 93 of 155
User Guide
Chapter 9: Populate Categories
2. In the Field Specifier list in the Select Field Specifier panel, select a field specifier.
The text in the Format, Examples, and Description fields provides guidance and information on
how to use the selected field specifier. For more information on field specifiers, see the HPE IDOL
Server Reference.
3. In the Set Arguments panel, enter argument values as required.
NOTE:
The Set Arguments panel adjusts for the type and number of arguments required for the
selected field specifier. For example, if the field specifier requires that you enter two values
as arguments to set upper and lower limits, there are two text fields.
4. In the Set Fields panel, use the Select Fields list to select the document fields to which you want
to apply the FieldText query filter. Click Add.
The selected document fields are shown in the Selected field.
NOTE:
You can select multiple document fields for the query filter.
5. In the Set Appending Boolean Operator panel, select an operator to append to the FieldText
query filter. Click Append.
The FieldText query filter is moved into the main Filter tab. If a query filter expression already
exists, HPE ICC appends the new query filter to it.
HPE IDOL Collaborative Classifier (11.1)
Page 94 of 155
User Guide
Chapter 9: Populate Categories
The Save Filter button flashes, prompting you to save the modified text.
When the FieldText filter is displayed in the Filter tab, you can test it. The procedure is similar to
testing a query (see Test a Query, below for more information). Documents that satisfy the query filter
are returned into the Review documents results area in the lower half of the tab area for inspection.
Test a Query
Before you save a query, you can test it to check that the results are correct.
To test a query
1. Select databases from the Database list.
2. Click Test Query
.
NOTE:
If the query text was modified but not saved before you clicked Test Query, a dialog box
opens, prompting you to save the query. If you do not save, the query is not performed.
HPE ICC runs the query, and the results appear in the Review documents results area.
3. Inspect the results to assess the query. There are two ways to show more document details:
l
Click Show/Hide summary of Documents
.
l
Double-click a document to launch a browser window with a three-pane view.
NOTE:
You must configure viewing to use this function. See Configure the Taxonomy Module, on
page 18.
4. If the result set is not satisfactory, revise the query text or filter field setting, and then retest it.
About Query Text and Filter History
The Query and Filter tabs enable you to view the history of all saved query texts or filters created by all
users for the category.
Whenever you change query or filter text, the Save Query or Save Filter button flashes, prompting you
to save the changes. If you try to test an unsaved query or filter, a dialog box opens asking you to save
the query text. You must save the text before you can test the query or filter.
Handle Unsaved Data
You are also prompted to handle unsaved data if you:
l
View the system summary System Summary.
l
Open a different taxonomy.
In both cases, the Confirm to Save dialog box opens.
If you choose not to handle the unsaved data, the view changes.
HPE IDOL Collaborative Classifier (11.1)
Page 95 of 155
User Guide
Chapter 9: Populate Categories
To handle unsaved data
1. In the Confirm to Save dialog, click Yes.
A dialog box opens.
All categories in the taxonomy that contain unsaved data are listed, along with their corresponding
category paths.
2. Select the category whose data you want to modify.
NOTE:
You can also select the When done, clear unsaved data check box. When selected,
HPE ICC deletes any unsaved data when you close the dialog box.
3. Click Find Category.
The category’s query or filter tab opens. Save, modify, or delete the unsaved text.
View and Retrieve Queries and Filters
You can retrieve a query text or filter expression to use or modify.
To view and retrieve a Query Text
1. In the Query tab, click Show History of Query Texts
.
The History of Query Texts dialog box opens.
2. The list of saved query texts is sorted in descending order by creation time. If you want to find
queries by a particular user, click the User ID column header to sort the list alphabetically. If you
want to search by query text, click the Query Text column header to sort the list alphabetically.
3. Select the saved query text that you want to use from the list.
The query text is shown in the Selected Record pane.
4. Click Copy to Clipboard.
5. Click Close.
6. Paste the copied query into the Query tab.
To view and retrieve a Filter
1. In the Filter tab, click Show History of Filters
.
The History of Filters dialog box opens.
2. The list of saved filters is sorted in descending order by creation time. If you want to search for a
particular user’s filter, click the User ID column header to sort the list alphabetically. If you want to
search by query text, click the Filter column header to sort the list alphabetically.
3. Select the saved filter that you want to use from the list. The filter is shown in the Selected
Record pane.
4. Click Copy to Clipboard.
5. Click Close.
6. Paste the copied query into the Filter tab.
HPE IDOL Collaborative Classifier (11.1)
Page 96 of 155
User Guide
Chapter 9: Populate Categories
Rules for Child Categories
If the selected category has child categories, you can import the rules from child categories and use
them as a basis for a query. Child categories appear in the taxonomy navigation pane.
To use a rule from a child category
1. In the Query tab, click Create Queries from Rules of Child Categories
.
NOTE:
If the selected category has no child categories, this icon is unavailable. If none of the child
categories have defined rules, an error message is displayed.
2. All query rules for child categories are imported into the Query tab. When multiple child category
rules exist, they are brought in as a single connected statement using the “OR” Boolean operator.
Manually Assign Documents
Use the Manual method to assign documents to a category manually. You must determine whether a
document’s subject matter qualifies it for inclusion in the category, rather than relying on automatic
selection performed according to a category rule. HPE recommends that you use the Manual method
only for categories for which you have expert knowledge of the subject matter.
Assign Documents to Accepted
When you assign a document to Accepted, it is modified in three ways:
l
It is immediately accepted for publishing.
l
It is given a score of 100.
l
A hand icon is displayed next to it in the document list.
To assign documents to Accepted manually
1. In the Manual search area, select the documents that you want to move to Accepted in the
Documents tab.
2. Click Move to Accepted
.
HPE ICC moves the documents to the Accepted tab on the Manual Action Results pane, and
adds them to the top of the document list on the Documents tab.
Assign Documents for Review
When you assign a document for review on the Documents tab, it is held there for review, and is not
published with the category.
HPE IDOL Collaborative Classifier (11.1)
Page 97 of 155
User Guide
Chapter 9: Populate Categories
To assign documents for review manually
1. In the Manual search area, select the documents that you want to move to review on the
Documents tab.
2. Click Move to To Be Reviewed
.
HPE ICC moves the documents to the To Be Reviewed tab on the Manual Action Results
pane, and adds them to the document list on the Documents tab. A hand icon is displayed next to
manually assigned documents.
Assign Documents to Excluded
To exclude documents from a category manually
1. In the Manual search area, select the documents that you want to exclude from the Documents
tab.
2. Click Move to Excluded
.
HPE ICC moves the documents to the Excluded tab on the Manual Action Results pane.
About the Manual Action Results Pane
When you manually assign documents, they appear under one of the three tabs in the Manual Action
Results pane. Use this pane to remove documents from each of the three assignments.
To remove a document from the Accepted, To Be Reviewed, or Excluded tab
1. Click the tab from which you want to remove documents.
2. Select the document that you want to remove from the tab. Click Remove From....
HPE ICC removes the selected document from the tab.
About Automatic Query Guidance
Use Automatic Query Guidance to obtain guidance in formulating query text that either forms the basis
of your category rule or helps you manage topics.
To use AQG, you must first enable it in the HPE IDOL configuration file. See Enable Automatic Query
Guidance (AQG), on page 18.
Enabling AQG alters the appearance of several screens in the Taxonomy pane:
l
l
In the Categorization tab, when Training is selected, the AQG button is located under the Search
bar and to the right of the Create Subtaxonomy button.
In the Categorization tab, when Query is selected, the AQG button is located in the upper left of the
document results section, and the Databases menu opens on the right.
AQG functionality is the same regardless of the screen from which you access it. You can perform
AQG from the search results if you are unsatisfied with the results of an existing query, or if you have
HPE IDOL Collaborative Classifier (11.1)
Page 98 of 155
User Guide
Chapter 9: Populate Categories
not yet conducted a query. You can also perform AQG from selected documents if you have already
conducted a query and want to refine your result set.
AQG enables HPE IDOL to summarize the results of a query or a number of manually selected
documents automatically. The summary clusters relevant terms from the documents or query results
and also provides a list of best terms, which are suggestions for alternative queries or training terms.
You can also try other query terms to see whether a more desirable set of documents is returned.
Run AQG From the Search Results
To run AQG from the search results
1. Click AQG.
The Automatic Query Guidance Wizard opens.
2. Click Next.
A query dialog box opens.
The Query box shows the query text.
If field text restrictions exist, they appear in the Field Text box. For more information, see Build a
Query Filter, on page 93.
3. Enter or modify the query and filter text, and then click Next.
A confirmation dialog box opens.
4. Click Create AQG.
The results dialog box opens (see About AQG Results, below).
Run AQG From Selected Documents
To use this option, you must first do one of the following:
l
Run a Query under the Categorization tab.
l
Run a Training search under the Categorization tab.
To run AQG from selected documents
1. Select one or more documents from the results of your query or search.
2. Click AQG.
The Automatic Query Guidance Wizard opens.
3. Select From Selected Documents, and then click Next.
A confirmation dialog box opens.
4. Click Create AQG.
The results dialog box opens (see About AQG Results, below).
About AQG Results
AQG clusters relevant terms in a tree structure, with the terms divided into nodes. At the top of the
dialog box is a collection of the three best concepts that AQG has identified.
HPE IDOL Collaborative Classifier (11.1)
Page 99 of 155
User Guide
Chapter 9: Populate Categories
NOTE:
You might have to enlarge the dialog box to view the best concepts if the text is too long.
The result set gives you an idea of how the information is organized. You can review the results and
use whatever terms you want for categorization or topic management purposes.
To use the AQG results
1. Select the terms that you want to use from the result set.
The terms appear automatically in the text box above the result list.
NOTE:
The best concepts do not automatically appear in the text box. If you want to use the best
concepts, you must locate and select them from the results list.
2. Click Copy Selected Terms.
All terms in the text box are copied to the clipboard.
3. Click Close.
4. Paste the terms into the target field (query or training text).
5. Run the query or training process.
HPE IDOL Collaborative Classifier (11.1)
Page 100 of 155
User Guide
Chapter 10: Populate Binary Categories
Chapter 10: Populate Binary Categories
This section describes how to populate binary categories.
•
•
•
Introduction
101
•
101
Binary Category Training Workflow
Train Binary Categories
101
•
•
Define Training Criteria
102
Train the Category
103
Binary Queries
103
Introduction
Binary categories function differently from regular categories.
Binary categories are all first-level children of the single Binary Taxonomy project. To be sure of
accurate query results, each category must receive both positive and negative training—that is, you
must define documents or text that you do want to associate with a query (positive training) as well as
documents or text that you do not want to associate with a query (negative training).
You can train binary categories on the Binary Actions tab. You can select training documents
manually, use all the documents in a specified directory, or enter specific training text.
NOTE:
The types of training do not have to be the same. For example, you can use documents for
positive training and a directory for negative training.
For more information about binary categories, see the HPE IDOL Server Administration Guide.
Binary Category Training Workflow
1. In the Binary Actions tab, select the Binary Training option.
2. Define some positive and negative training documents, directories, or text.
3. Train the category.
4. Select the Binary Query option.
5. Enter a query to test the category training.
6. Refine the training as required.
The following sections provide more details on training options.
Train Binary Categories
Binary category training is a two-stage process. You must first define positive and negative training
criteria. After you define your training material, you can train the category and view training terms and
HPE IDOL Collaborative Classifier (11.1)
Page 101 of 155
User Guide
Chapter 10: Populate Binary Categories
weights.
Define Training Criteria
There are three ways to define training criteria:
l
Use individual documents. See Select Training Documents, below.
l
Use training text. See Enter Training Text, below.
l
Use a directory. See Select Training Directories, on the next page.
You can use any combination of the three methods for positive and negative training.
Select Training Documents
If you want to select specific documents to use for positive or negative training, you must first search
one or more HPE IDOL databases. The search process in the binary category interface is the same as
in the regular taxonomy interface. For more details about searching, see Search a Database, on page
78.
For binary categories, it might be beneficial to perform multiple searches to identify positive and
negative training documents.
For example, if you had a binary category called “Japanese cars,” you could search for “Nissan,”
“Toyota,” and “Honda,” and then use selected documents for positive training, because these are all
Japanese brands. You could then search for “Ford,” “GM,” and “Chrysler,” and use appropriate
documents for negative training, because these brands are not Japanese.
To select training documents
1. Perform a search.
The search results appear in the document list.
2. Select the documents that you want to use for training.
3. For positive training click
, or for negative training click
.
The documents appear in the result list in the Training Docs tab.
To remove documents from the training documents list
1. Select the documents that you want to remove from the training list.
2. Click Remove from Training
.
Enter Training Text
The training text procedure for binary categories is similar to the procedure for regular categories.
However, you can enter both positive and negative training text. For example, if the training category
were “Asian cars,” you could enter “Nissan,” “Toyota,” and “Honda” as positive training text and “Ford,”
“GM,” and “Chrysler” as negative training text.
HPE IDOL Collaborative Classifier (11.1)
Page 102 of 155
User Guide
Chapter 10: Populate Binary Categories
To enter training text
1. Select the Training Text tab.
2. Enter either positive training text, negative training text, or both.
3. Click Save.
Select Training Directories
Instead of selecting training documents manually, you can use all the documents in a specified
directory for either positive or negative training.
To select directories for training
1. Select the Training Dir tab.
2. Click
to browse to a positive training directory, or
directory. In the dialog box, click Save.
to browse to a negative training
OR
Enter a full directory path in the Positive text box, Negative text box, or both.
3. Click Save.
Train the Category
After you have defined some positive and negative training, you can train the category and view training
terms and weights.
To train the category
l
Select any Binary Training tab, and then click Train.
View Terms and Weights
After you train the category, the Terms and Weights tab is populated. However, unlike in a nonbinary
category, you cannot modify these terms and weights manually.
Binary Queries
After you train a binary category, you can perform a binary query against the category. Although the
query interface is similar to a regular nonbinary category query, the purpose and the results are
different. A binary query determines whether the query text belongs in the category.
When you run a binary query, HPE IDOL Collaborative Classifier returns either a POSITIVE result,
which indicates that the query text belongs in the category, or a NEGATIVE result, which indicates that
the text does not belong in the category. There is also a corresponding score (from 0 to 1) that indicates
how well the text fits into the binary category.
For more information on binary category queries, see the HPE IDOL Server Reference.
HPE IDOL Collaborative Classifier (11.1)
Page 103 of 155
User Guide
Chapter 10: Populate Binary Categories
To run a binary category
1. On the Binary Actions tab, select the Binary Query option.
2. In the text box, enter the query text. Click Save Query.
3. Click Run Binary Query
.
The query result and score appear beneath the text box, under Binary Query Response Data.
HPE IDOL Collaborative Classifier (11.1)
Page 104 of 155
User Guide
Chapter 11: Manage Documents
Chapter 11: Manage Documents
This section describes how to manage a document set in the Taxonomy module.
•
•
•
•
•
About Document Management
105
Configure the Document List
106
•
•
•
•
Add or Remove Columns in the Document List
107
Change the Document List Output
108
Change the Documents Shown
108
Paging Managers
109
About the Document Distribution Graph
109
•
•
Read the Graph
110
Change the Threshold
111
Manipulate Documents in the List
111
•
•
•
•
•
•
Move Documents into Another View
112
Remove Approved Documents from a Category
112
Change a Document Ranking
112
Recommend Documents to Other Categories
113
Revert a Manual Action
113
Mark Documents for Training
113
Filter the Category
114
About Document Management
The Documents tab contains two secondary tabs.
HPE IDOL Collaborative Classifier (11.1)
Page 105 of 155
User Guide
Chapter 11: Manage Documents
The Document List shows a list of the documents associated with a category, and the Chart shows a
document distribution graph for that set of documents. These tabs enable you to review and manage
the documents associated with a category.
You can manipulate documents in the following ways.
l
Move them to the Accepted or To Be Reviewed areas
l
Remove them from the document list
l
Recommend them to another category
l
Mark them relevant for training (see Mark Documents for Training, on page 81)
Configure the Document List
You can view documents placed into the category on the Document List secondary tab on the
Documents tab. The Document List secondary tab is configurable, enabling you to view the
document list with different headings and information. For example, you can change the document list
to view the Title, Location, Score, and URL for each document.
HPE IDOL Collaborative Classifier (11.1)
Page 106 of 155
User Guide
Chapter 11: Manage Documents
Add or Remove Columns in the Document List
The Available list includes several headings.
To change the text header columns displayed in the document list, right-click the heading area. Use the
following procedures to edit the column display.
To add a heading
1. In the Available list, select the heading that you want to add.
2. Click the right arrow (→) to move the new heading to the Selected list.
To remove a heading
1. In the Selected list, select the heading that you want to remove.
2. Click X.
To move a heading up or down the list
1. In the Selected list, select the heading that you want to move.
2. Use the up and down arrows to move a selected item in the list, changing its display order in the
table.
To view the updated list
l
Click outside the dialog box.
NOTE:
The HPE IDOL Collaborative Classifier category contains the headings specific to HPE ICC,
such as document view, action, and training status. The following procedures assume that all
HPE IDOL Collaborative Classifier (11.1)
Page 107 of 155
User Guide
Chapter 11: Manage Documents
HPE ICC headings are displayed.
Change the Document List Output
Use the following user interface elements to alter the documents list:
l
Fit columns to width. Fit the columns to the width of the pane.
l
Show or Hide summary of documents. Show expanded document summaries.
l
Database list. Select which databases to use for document retrieval.
l
Extended document view. Double-click a document to launch a browser window with a three-pane
view giving extended details of the document.
NOTE:
You must configure viewing to use this function. See Installation and Configuration, on page
15.
Change the Documents Shown
You can view different documents based on the document view and action in the document list.
View Documents Added Manually or Automatically
To view documents added manually or automatically
1. In the column header, click Action
.
2. Select or deselect Show Manual or Show Automatic.
HPE ICC refreshes the document list.
View Documents in Different Views
You can filter the document list to show documents of other statuses.
To view documents in different views
1. In the column header, click View
.
2. Select Show Accepted, Show To Be Reviewed, or Show Excluded.
HPE ICC refreshes the document list.
HPE IDOL Collaborative Classifier (11.1)
Page 108 of 155
User Guide
Chapter 11: Manage Documents
Paging Managers
By default, you can view only the first page of documents in the Document List, usually the first 100
hits. To view more than one page, you can enable paging managers (see Enable Paging Managers, on
page 20). When you enable paging managers, the Document List appearance changes slightly.
In addition, there are several changes to functionality:
l
The maximum number of results that appear is limited by the MaxResults HPE IDOL configuration
parameter. Note that the MaxResults value might be smaller than the total hits in the document list.
If the Total Hits value is greater than the Max Results value, both totals are reported above the
document list; otherwise, only the Total Hits value is displayed.
For more information on the MaxResults parameter, see the HPE IDOL Server Reference.
l
l
The Filter field does not appear above the document list. See Filter the Category, on page 114.
Any documents that were manually moved are displayed only on the first page. The number of
manually assigned documents is added to the maximum number of documents per page.
For example, if you set the maximum number of documents to 100, and there are 9 manually
assigned documents, the first page shows 109 documents.
l
l
l
Manual actions from the document list panel can be used only on the first page. To manually change
document ranking, users must use the Manual categorization method. See Manipulate Documents
in the List, on page 111.
The Chart shows the distribution of documents from the current page. If you change the chart
thresholds and return to the document list, the first page is displayed. See Change the Threshold, on
page 111.
When you select a different node that has documents in the Document List, the first page is
displayed.
About the Document Distribution Graph
You can view the document distribution graph from the Chart secondary tab on the Documents tab.
The graph displays the relevance of documents to a category. Documents are distributed along the
graph depending on their score. You can set and alter thresholds to affect how documents are placed in
the category.
HPE IDOL Collaborative Classifier (11.1)
Page 109 of 155
User Guide
Chapter 11: Manage Documents
Read the Graph
The Document Distribution Graph displays a rough estimation of documents across a graph according
to where they fall within the lower and upper acceptance thresholds.
The elements of the document distribution graph include the following:
l
l
Red arrowheads show where the Accepted or To Be Reviewed thresholds are set. See Change the
Threshold, on the next page.
Tooltips on the colored areas that indicate:
o
Red: excluded
o
Yellow: to be reviewed
o
Green: accepted
HPE IDOL Collaborative Classifier (11.1)
Page 110 of 155
User Guide
Chapter 11: Manage Documents
l
Green dots indicate documents that are manually marked as relevant for the category.
NOTE:
The locations of the green dots provides information about the quality of the category training.
For example, a number of green dots in the red area (rejected) portion of the graph might
indicate that the documents are scoring too low to be considered good training documents.
l
Graph dips and peaks indicate approximately where the documents fall on the distribution graph with
respect to scores. The Peak number of documents at the top-left corner refers to the peak of the
graph.
NOTE:
You can view score information for training documents on the Training tab.
Change the Threshold
You can change the document distribution graph thresholds for a category to enable fewer or more
documents to be added automatically. For example, you might want to move thresholds to include or
exclude document peaks.
To change the threshold
1. Slide the threshold bars to move the Accepted or To Be Reviewed thresholds.
A confirmation dialog box opens, in which you can manually adjust the values.
l
l
Auto-Publish Threshold. Set the threshold for documents to be moved directly into
Accepted for publication.
Inbox Threshold. Set the threshold for documents to be reviewed. HPE ICC excludes
documents below this threshold from the category view.
2. Click Yes.
HPE ICC refreshes the document distribution graph with the new values.
Manipulate Documents in the List
You can perform the following actions with documents in the Document List tab:
l
Move Documents into Another View, on the next page.
l
Remove Approved Documents from a Category, on the next page.
l
Change a Document Ranking, on the next page.
l
Recommend Documents to Other Categories, on page 113.
l
Revert a Manual Action, on page 113.
l
Mark Documents for Training, on page 113.
You can perform many of the actions on multiple documents.
Before you can manipulate any documents in the list, you must lock the taxonomy category.
HPE IDOL Collaborative Classifier (11.1)
Page 111 of 155
User Guide
Chapter 11: Manage Documents
NOTE:
If paging managers are enabled, you can perform manual actions only on the first page of
documents in the list; you can modify other documents only with the Manual categorization
method. See Paging Managers, on page 109.
Move Documents into Another View
To move a document or documents between views to change the status
1. Select the document or documents that you want to move.
2. Click the appropriate view button to move the documents. The views available are:
Accepted. Documents are placed as to be published to the
category node.
Document Review. You can review documents before they are
placed in Accepted or removed from the list.
The view icon changes accordingly.
Remove Approved Documents from a Category
To remove a document or documents from your category
1. Select the document or documents you want to remove.
2. Click Remove from View
.
The list refreshes and the documents disappear from the category. You can view the documents in
the Excluded view.
Change a Document Ranking
To change the document ranking of a manually ranked document
1. Select the document to change.
NOTE:
You can change only one document manually at a time.
2. Click Change Rank for Manual Staging Document
.
The Change Rank dialog box opens.
3. Select the rank number from the list. Click OK.
The document list refreshes with the new ranking.
HPE IDOL Collaborative Classifier (11.1)
Page 112 of 155
User Guide
Chapter 11: Manage Documents
Recommend Documents to Other Categories
To recommend documents to other categories
NOTE:
Users with reviewer privileges can recommend documents to another category if they have
editor privileges on the other category.
1. Select the document that you want to recommend.
2. Click Recommend to another category
.
The Select Category dialog box opens.
3. Select the category to which you want to recommend the document. Click OK.
The document opens as a manually ranked document in the To Be Reviewed list for the
recommended category.
Revert a Manual Action
You can revert a manual action.
For example, if you manually move a document from To Be Reviewed into Accepted view, but then
revert that action, the document is returned to To Be Reviewed.
To revert a manual action to a document
1. Select the manual document.
2. Click Clear Manual Action
.
The document is returned to its original state.
Mark Documents for Training
To mark documents as relevant for training
1. Select one or more documents.
2. Click Relevant for Training
.
The documents are marked as relevant for training.
Although the documents are marked relevant for training, you must still train the category in the
Training Documents area. See Define Category Rules by Using Training Documents, on page 81.
HPE IDOL Collaborative Classifier (11.1)
Page 113 of 155
User Guide
Chapter 11: Manage Documents
Filter the Category
You can add a filter to the category. Returned documents must satisfy the filter criteria as well as the
category rule.
NOTE:
You cannot add filters if paging managers are enabled. See Paging Managers, on page 109.
To add a filter to the category
l
Enter the filter text in the Filter text box on the Documents List tab, or click the arrow to the right of
the Filter text box to select a saved filter from a list.
HPE IDOL Collaborative Classifier (11.1)
Page 114 of 155
User Guide
Chapter 12: Review Categories
Chapter 12: Review Categories
This section discusses the tasks performed by users with reviewer privileges.
•
•
•
•
Reviewer Tasks
115
Verify a Category Rule
115
Add Review Comments to a Category
118
Generate Category Reports
118
Reviewer Tasks
Reviewers can perform the following actions:
l
Open and close taxonomies and resource taxonomies
l
View the draft, filtered draft, and published views of taxonomies
l
Lock and unlock categories
l
Add review comments to a category in the Category Details tab
l
Perform searches from the Training tab
l
Verify a category rule
l
Create custom filters
l
Request publication of a category
l
Recommend a document to another category
l
Generate category reports
This section explains the reviewer tasks of verifying a category rule, adding review comments to a
category, and generating category reports.
Verify a Category Rule
Taxonomy reviewers can use the QA tab to verify the quality of a category using a set of benchmark
documents.
HPE IDOL Collaborative Classifier (11.1)
Page 115 of 155
User Guide
Chapter 12: Review Categories
To verify a category rule
1. Select the category in the Taxonomy pane whose rule you want to test.
2. Search for documents. See Search for Documents, below.
3. Examine the search result documents that appear in the upper half of the tab.
4. Choose documents that are good examples of the subject matter that the category must contain,
and tag them for benchmarking. See View Documents and Marking for Benchmark, below.
5. When you have a satisfactory set of benchmark documents, evaluate the rule. See Evaluate the
Category Rule, on the next page.
6. Use the evaluation to assess the performance of the category rule against the benchmark.
Search for Documents
Use the Select Benchmark Documents area on the QA tab to perform a search. See Search a
Database, on page 78.
View Documents and Marking for Benchmark
Examine the documents returned as the search result set. There are two ways you can view more
details of these documents to assist your selection:
l
Click Show/Hide summary of Documents
l
Double-click a document to launch a browser window with a three-pane view.
HPE IDOL Collaborative Classifier (11.1)
.
Page 116 of 155
User Guide
Chapter 12: Review Categories
NOTE:
You must configure viewing to use this function. See Configure the Taxonomy Module, on page
18.
To select documents as relevant for benchmark
1. Select the documents that you want to use as benchmark documents.
2. Click Relevant For Benchmark
.
The documents appear in the lower half of the QA tab under Benchmark Results on the
Benchmark Documents tab.
Remove Documents from Benchmark
To remove benchmark documents
1. In the Benchmark Documents tab, select the documents that you want to remove.
2. Click Remove from Benchmark
.
HPE ICC removes the selected documents from the benchmark set.
Evaluate the Category Rule
After you assemble a benchmark set of documents, you can evaluate the category rule against the
benchmark.
The category rule is evaluated against the selected databases. The documents returned for the rule that
score above the category’s acceptance threshold appear in the Precision and Recall tab.
In the Precision and Recall tab, all documents returned by the category rule that match a benchmark
document are checked as matches, and the evaluation is divided into separate measures.
Retrieved
The number of documents returned based on the category rule and that score above
the category’s acceptance threshold.
Benchmarked The number of documents in the benchmark set.
Matched
The number of documents in the rule return set that match a benchmark document.
Precision
The number of matched documents divided by the total number of documents
returned.
Recall
The number of matched documents divided by the total number of benchmarked
documents.
Overall
The weighted harmonic mean of precision and recall. The traditional F-Measure,
because precision and recall are evenly weighted, is
HPE IDOL Collaborative Classifier (11.1)
Page 117 of 155
User Guide
Chapter 12: Review Categories
To evaluate a category rule
l
Click Evaluate.
Add Review Comments to a Category
To add review comments to individual categories
1. Select the category to which you want to add comments.
2. Lock the category.
3. Click the Category Details tab.
4. Click the Status secondary tab.
5. Enter comments into the Reviewer Notes text box.
6. Click Save Changes.
NOTE:
You can inspect a history of reviewer notes on this tab and retrieve saved reviewer notes. See
Add Reviewer Notes, on page 63.
Generate Category Reports
The reporting function is intended for reviewing and auditing a category. You can generate a report on
the basis of several features of a category:
l
Total Hits
l
Current Role Status
l
Benchmark
l
Query
l
Reviewer Notes
l
Filter
To generate a report
1. In the Taxonomy pane, click the category on which you want to report.
2. Click Report.
The Generate Report dialog box opens.
3. Select the type of report that you want.
NOTE:
To generate a Total Hits report, click Go to verify whether the data is in sync. If it is, the
Total Hits option is enabled. If the data is not in sync, make sure that the entire taxonomy
tree is published, and then try again.
4. Under Select Scope of Report, select a report of the single selected category node or the entire
node subtree.
HPE IDOL Collaborative Classifier (11.1)
Page 118 of 155
User Guide
Chapter 12: Review Categories
5. Under Save Report as, enter or browse to an XLS or CSV file to which you want to save the
report.
If you enter or select an existing file, a confirmation dialog box opens. Click Yes to overwrite the
file or No to enter a different file name.
6. Click OK.
A report file is created and saved in the specified directory.
HPE IDOL Collaborative Classifier (11.1)
Page 119 of 155
User Guide
Chapter 12: Review Categories
HPE IDOL Collaborative Classifier (11.1)
Page 120 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
Chapter 13: Publish and Export a Taxonomy
This section describes how to publish and export taxonomies.
•
•
•
•
About Publishing
121
Publish Categories
121
Publish Binary Categories
123
Export a Taxonomy
124
About Publishing
After you populate a category, you can publish it. Before a category is published, it is available only
internally to the HPE IDOL Collaborative Classifier. Publishing a category makes it available to the
HPE IDOL Server.
Publishing a category depends on user roles. If you are assigned a Publisher user role for a category in
a taxonomy, you can perform all the actions that a Reviewer or an Editor can perform, and you can also
publish the category.
User Roles
l
Reviewer. Cannot publish and the Publish button is grayed-out.
l
Editor. Can request publication; user sees an active Request to Publish button.
l
Publisher. Can publish; user sees an active Publish button.
HPE ICC also allows you to export a taxonomy to an XML file.
Publish Categories
To publish a category, you must have publisher privileges for that category, and the parent category
must already be published.
Publish a Category
To publish a category to the host server
1. In the File menu, select Open Taxonomy.
2. Select the category in the taxonomy pane.
3. Click Publish.
l
If there is a subtree of child categories under the category node, a confirmation dialog box
opens. Select This node only to publish a single node, or Entire subtree to publish the entire
subtree. Click OK.
HPE IDOL Collaborative Classifier (11.1)
Page 121 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
NOTE:
You can publish a subtree only if you have publishing privileges on all categories.
l
If there are no child categories, a dialog box opens asking you to confirm the publishing
request. Click Yes.
NOTE:
If the parent category has not been published, an error message opens in a dialog box.
Click OK to cancel the publishing process.
HPE ICC publishes the category on the host server, and a green circle is displayed on its folder
icon in the Taxonomy pane.
Publish a Root Category
When you publish a root category (the category that sits at the top level of the category hierarchy), you
can publish it under a different parent category. You can publish root categories to the host server or to
a remote server.
To publish a root category
1. Select the root category in the Taxonomy pane.
2. Click Publish.
The Publish Taxonomy dialog box opens.
3. Select This node only to publish a single node, or Entire subtree to publish the entire subtree.
4. Click Verify to verify the connection to the server. If the verification fails, make sure that the host
and port information is correct, and then try again.
5. To select the parent category under which the root category is published, click Browse.
The Select Category dialog box opens.
6. Select a parent category. Click OK.
The Select Category dialog box closes and the selected parent category is shown on the Publish
Taxonomy dialog box in the Selected category name field.
7. Make sure that the correct parent category is selected. If not, click Browse and repeat Step 5 to
Step 7.
8. Click Publish.
HPE ICC publishes the root category (or the root category and subtree) on the host server under
the parent category.
Publish to a Remote Server
You can publish a root category to a different server. For example, if you have a staging server and a
production server, you could publish a category to a staging server first. You could then remotely
publish to the production server at the appropriate time.
HPE IDOL Collaborative Classifier (11.1)
Page 122 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
NOTE:
The entire category tree must already be published before you can publish to a remote server.
To publish a category to a remote server
1. Select the root category in the Taxonomy pane.
2. Click Publish.
The Publish Taxonomy dialog box opens.
3. Click Go to verify whether the data is in sync. If it is not, an error message opens. Make sure that
the entire taxonomy is already published to the local server.
When the data is in sync, the Remote Publish option is enabled and selected.
NOTE:
When Remote Publish is selected, the Entire subtree option is automatically selected;
you cannot publish only a single node to a remote server.
4. Enter the server name and port to which you want to publish the taxonomy.
5. Click Verify to verify the connection to the server. If the verification fails, make sure that the host
and port information is correct, and then try again.
6. Under Select parent category, click Browse.
The Select Category dialog box opens.
7. Select a parent category. Click OK.
The Select Category dialog box closes, and the selected parent category is shown in the Publish
Taxonomy dialog box in the Selected category name field.
8. Make sure that the correct parent category is selected. If not, click Browse and repeat Step 6 to
Step 7.
9. Click Publish.
HPE ICC publishes the taxonomy on the remote server.
Publish Binary Categories
To publish a binary category, you must have publisher privileges for that category. You can publish
binary categories to the host server or to a remote server.
Publish a Binary Category
You can publish binary categories individually because they are not organized in the same way as
regular categories (see About Binary Taxonomy Authoring, on page 69). You cannot publish the Binary
Taxonomy root node.
To publish a binary category
1. Select the binary category in the taxonomy pane.
2. Click Publish.
The Publish Binary Category dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 123 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
3. Click Verify to verify the connection to the server. If the verification fails, make sure that the host
and port information is correct, and then try again.
4. Click Publish.
HPE ICC publishes the binary category on the host server, and the category icon turns green.
Publish to a Remote Server
You can publish a binary category to a different server. For example, if you have a staging server and a
production server, you could publish a category to the staging server first. You could then remotely
publish to the production server at the appropriate time.
To publish a binary category to a remote server
1. Select the root category in the Taxonomy pane.
2. Click Publish.
The Publish Taxonomy dialog box opens.
3. Select the Remote Publish check box. The host and port information becomes editable.
4. Enter the target Category server and Search server names and ports to which you want to publish
the taxonomy.
5. Click Verify to verify the connection to the server. HPE IDOL Collaborative Classifier checks
whether the name of the selected binary taxonomy already exists on the target server. If it does,
the options under Handling Remote Name Conflict are enabled, and you can resolve the
conflict in one of two ways:
l
l
Change Current Name. Change the name of the binary category that you want to publish.
(Click Browse to view the complete list of binary categories on the target server.)
Overwrite Existing. Replace the category that already exists on the target server. This option
is selected by default.
6. Click Publish.
HPE ICC publishes the category on the remote server.
NOTE:
When you publish a binary category on a remote server, the category icon does not turn
green.
Export a Taxonomy
You can export a taxonomy to an XML file format to make it available to user client applications.
To export a taxonomy
1. In the Taxonomy pane, select the category you want to export.
2. Click Export.
The Save Taxonomy XML File dialog box opens.
3. Browse to the directory in which you want to save the taxonomy XML file.
HPE IDOL Collaborative Classifier (11.1)
Page 124 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
4. Enter a name for the taxonomy XML file in the File name field.
5. Click Save.
HPE ICC creates the taxonomy XML file and saves it in the chosen directory.
HPE IDOL Collaborative Classifier (11.1)
Page 125 of 155
User Guide
Chapter 13: Publish and Export a Taxonomy
HPE IDOL Collaborative Classifier (11.1)
Page 126 of 155
User Guide
Part 3: Applications
This section describes how to use the Applications module.
l
Create Applications
l
Modify Applications
HPE IDOL Collaborative Classifier (11.1)
Page 127 of 155
User Guide
HPE IDOL Collaborative Classifier (11.1)
Page 128 of 155
User Guide
Chapter 14: Create Applications
Chapter 14: Create Applications
This section describes how to create and modify applications.
•
•
•
•
About Applications
129
Activate the Applications Module
130
Work with Applications
130
•
•
Open and Close Applications
130
Refresh an Application
130
Add Applications
131
•
•
Create a New Configuration
131
Register an Application
132
About Applications
You can use the Application module to create and modify applications. With this module, you can use
the Component Framework to manage applications that work with HPE IDOL. Datasheets, catalogs,
and other content can be structured to create specific navigation paths for the users.
Figure 4: Basic Application module workflow
HPE IDOL Collaborative Classifier (11.1)
Page 129 of 155
User Guide
Chapter 14: Create Applications
Activate the Applications Module
The applications module is disabled by default. You must enable it in the acc.properties
configuration file.
To configure HPE IDOL Collaborative Classifier to display the Applications module
1. Go to <installDir>\appServer\webapps\icc\WEB-INF\classes\config.
2. Open acc.properties.
3. Set the module.applications.enable parameter to True.
module.applications.enable=True
4. Restart Tomcat.
Work with Applications
This section discusses the Applications module interface and how you can work with it.
Open and Close Applications
You can open and close applications by using the Application Module. You can also move between
open applications by using the Window menu.
To open an application from the menu
1. In the File menu, select Open, and then select Application.
2. Select the application that you want to open from the Open Application dialog box. Click OK.
The application opens.
To open an application from the Application Summary pane
1. In the Summary window, select the application that you want to open from the application list in
the Application pane.
2. Click Open.
The application opens.
NOTE:
You can also double-click the application to open it.
To close an application
l
In the File menu, select Close.
Refresh an Application
Refresh your application to return the draft version to the most recently published version.
HPE IDOL Collaborative Classifier (11.1)
Page 130 of 155
User Guide
Chapter 14: Create Applications
To refresh an application
1. Click Refresh.
A confirmation dialog box opens.
2. Click Yes to restore your last published version.
Click No to continue working in your application.
Add Applications
You can use the Add New Application wizard to:
l
Create a new configuration.
l
Register an existing configuration.
Create a New Configuration
Use the Add New Application wizard to set the basic configurations of your new application.
To create a new application
1. In the File menu, select New, and then select Application.
The New Application wizard opens.
2. Select Create new configuration. Click Next.
The Enter Alias, Host and Port for the Application wizard screen opens.
3. Enter the new application’s name (alias), host, and port. Click Verify.
If the connection to the server is valid, Valid is displayed on the right of the screen.
4. Click Next.
The Web Application Location wizard screen opens.
5. Enter the file path to your web application. Click Verify to validate the path.
NOTE:
Create the application directory on the file system before you register it in HPE IDOL
Collaborative Classifier. The path to the application directory must include the name of the
directory, but not the config directory. A trailing slash is optional.
6. After the path is validated, click Next.
The Application Preview URL wizard screen opens.
(Optional) Enter the URL to use to preview your application.
NOTE:
The URL must point to the application directory. For example, if your application is in your
webapps directory on the same machine as your HPE ICC installation, your URL might be:
http://<hostname>:9990/<application_directory>.
7. Click Next.
HPE IDOL Collaborative Classifier (11.1)
Page 131 of 155
User Guide
Chapter 14: Create Applications
The Application Configuration File Name wizard screen opens.
8. Enter a name for the application configuration file. Click Next.
The Confirm actions wizard screen opens.
NOTE:
Click Finish to create the application configuration using the defaults.
9. Confirm your settings, and then click Create Application.
The creation process runs in the wizard.
10. Click Open New Application after the creation process is complete.
The new application opens in HPE ICC.
Register an Application
If you already have an application that you want to configure through HPE ICC, you can register it in the
Application module.
To register an application
1. In the File menu, select New, and then select Application.
The New Application wizard screen opens.
2. Select Register existing configuration. Click Next.
The Enter an Alias for the Application wizard screen opens.
3. Enter the name (alias) of the application. Click Next.
The Web Application Location wizard screen opens.
4. Enter the path to the web application.
5. Click Verify to validate the path.
NOTE:
The path must point to the top level of the application directory, not the config directory.
6. After the path is validated, click Next.
The Application Preview URL wizard screen opens.
7. (Optional) Enter the URL to use to preview your application.
NOTE:
The URL must point into the application directory. For example, if your application is in
your webapps directory on the same machine as your HPE ICC installation, your URL
might be:
http://<servername>:9990/<application_directory>.
8. Click Next.
The Existing Application Configuration Files wizard screen opens.
9. Select one of the valid configuration files from the list. Click Next.
The Confirm Actions wizard screen opens.
HPE IDOL Collaborative Classifier (11.1)
Page 132 of 155
User Guide
Chapter 14: Create Applications
NOTE:
Click Finish to create the application configuration using default values.
10. Confirm your settings. Click Create Application.
The creation process runs in the wizard.
11. After the creation process is complete, click Open New Application.
The new application opens.
HPE IDOL Collaborative Classifier (11.1)
Page 133 of 155
User Guide
Chapter 14: Create Applications
HPE IDOL Collaborative Classifier (11.1)
Page 134 of 155
User Guide
Chapter 15: Modify Applications
Chapter 15: Modify Applications
This section describes how to modify your applications.
•
•
•
•
•
•
•
Add Contexts
135
Add Components
136
Delete Contexts or Components
138
Edit Contexts or Components
138
View the XML
139
Preview the Application
139
Publish an Application
139
Add Contexts
Contexts specify the connection between an HPE Component Framework service, its components,
and other information required by the Component Framework to manage components. There are
different kinds of contexts that represent different parts of the Component Framework.
A context defines:
l
The host and port of an HPE Component Framework Server.
NOTE:
The Component Framework might be running in a brokered environment, in which case the
context is the broker’s host and port; otherwise, the context is an HPE Component
Framework Server’s host and port.
l
A type of index, such as a collection or parametric index.
You can associate a component with one or more contexts. You can share a component across
contexts; thus, you need to define a component only once and associate it with each context in which
you use the component. The primary context specifies the source of a component’s XML or HTML
data.
To add a context to an application
1. Select the application root.
2. Click Edit to lock the application.
3. Click Add Context.
The Add Context dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 135 of 155
User Guide
Chapter 15: Modify Applications
4. Select the context type from the list.
5. Enter the name of the new context.
6. Click OK.
HPE ICC adds the new context to the application.
Add Components
You can add components to your application context by creating a new component or sharing an
existing component.
Add a New Component
To add a new component
1. Select the context element to add a component to.
2. Click Edit to lock the application.
3. Click Add Component.
The Add Component dialog box opens.
HPE IDOL Collaborative Classifier (11.1)
Page 136 of 155
User Guide
Chapter 15: Modify Applications
4. Select New Component.
5. Select the component type from the list.
6. Enter the name of the component type.
7. Click OK. HPE ICC adds the component to the application context.
Share a Component
To copy a component from another context
1. Select the context element.
2. Click Add Component.
The Add Component dialog box opens.
3. Select the Share Existing option.
4. Select a component from the list.
5. Click OK.
HPE ICC adds the shared component to the context.
HPE IDOL Collaborative Classifier (11.1)
Page 137 of 155
User Guide
Chapter 15: Modify Applications
Delete Contexts or Components
To delete entire contexts or components from a context
1. Select the context or component that you want to delete.
2. Click Delete.
A confirmation dialog box opens.
3. Click Yes.
HPE ICC deletes the context or component.
Edit Contexts or Components
This section discusses how to edit contexts or components in the user interface.
To edit contexts or components
1. Select the context or component to edit.
The available context or component fields are shown next to the application pane.
2. Edit the fields as required.
See Edit Fields, below for suggestions about the application field user interface.
3. Publish the changes.
See Publish an Application, on the next page for information about publishing your application
changes.
NOTE:
Fields that correspond to the HPE IDOL or HPE IDOL Collaborative Classifier system are
not automatically updated. To update these fields, click Refresh in the Applications
toolbar.
Edit Fields
Use the application fields to specify attributes and view parameters. The field user interface has the
following behavior:
l
l
Required fields are marked with a red asterisk.
Select or clear fields by using the check boxes or the plus and minus buttons next to the fields. The
plus or minus buttons expand or collapse the related field.
l
Set field values in lists and text boxes.
l
If you enter data that is not valid in a field, several user interface actions might occur:
o
The context or component icon displays a red x.
o
The text box changes to a pale red color.
HPE IDOL Collaborative Classifier (11.1)
Page 138 of 155
User Guide
Chapter 15: Modify Applications
l
o
A validation error opens in the message panel.
o
A dialog box opens with a suggestion about creating a valid field.
Some fields are dependent on the value set for another field, affecting the available lists or valid text
entries.
View the XML
To view the application XML
1. Click Show XML.
A window displays the current XML configuration.
Preview the Application
You can review the changes to your application to check that the published results are satisfactory. To
review the application, you must add a preview URL during the creation process. If you have
Administrator privileges, you can also add or edit this URL from the Administration window. See Edit an
Application URL, on page 37.
l
Click Preview.
Your default browser displays the application.
Publish an Application
After you change the application, you must publish the changes.
To publish an application
1. Make sure that the changes are shown in the application configuration pane.
2. Click Edit to lock the application.
3. Click Publish.
The changes are published.
NOTE:
If the application validation discovers an error in the XML, a confirmation dialog box opens.
Click Yes to publish, or No to reject the changes.
HPE IDOL Collaborative Classifier (11.1)
Page 139 of 155
User Guide
Chapter 15: Modify Applications
HPE IDOL Collaborative Classifier (11.1)
Page 140 of 155
User Guide
Part 4: Appendixes
This section includes the following appendixes:
l
Workflow Example
l
HPE Category XML Format
HPE IDOL Collaborative Classifier (11.1)
Page 141 of 155
User Guide
HPE IDOL Collaborative Classifier (11.1)
Page 142 of 155
User Guide
Appendix A: Workflow Example
Appendix A: Workflow Example
This section includes a possible workflow at Company XYZ using the Taxonomy and Application
modules. This example shows many of the recommended practices for using HPE IDOL Collaborative
Classifier.
Company XYZ wants to create an application that can display their products, product documentation,
and product experts in each company location. They have a collection of documents to use initially.
Company XYZ has installed HPE ICC on a single host that business users modify, and that end users
use to search and browse.
HPE ICC workflows are dependent on user permissions, so Company XYZ assigns user types to HPE
ICC for administrator, developer, and librarian.
The administrator is responsible for maintaining the resources, including deleting resources, polling
collections, and assigning user permissions to the modules. The administrator for Company XYZ
assigns creator privileges for each module to an individual user of each type. These users are then
responsible for creating resources in the modules. When resources have been created, either the
resource owners or the administrators can assign additional roles to the individual taxonomies:
l
l
l
l
The editor is responsible for the resource structure and the contents. The editor can add documents
to categories, move categories within the taxonomy, and update the category as needed.
After a resource such as a taxonomy is created, a reviewer can be assigned. The reviewer reviews
the category documents and provides comments to either the editor or publisher of that category.
After the editor and reviewer have created and approved a category, the publisher can publish the
category and its contents to the public domain.
The librarian is assigned certain taxonomies to create for use. These taxonomies become the draft
taxonomies to which the editors, reviewers, and publishers can add content.
The publisher (who is usually the primary librarian for a category) creates subcategories in the
taxonomy, and assigns the editor and reviewer roles to the categories.
The editor can open a taxonomy, search for documents or create topics in the topic editor, add those
documents to the document list, and request either a review or publication of the category and content.
The editor’s daily tasks for a taxonomy might include running the workspace filter, which shows
editable and locked categories, and categories that contain unpublished changes. Also, the editor might
regularly look at documents to be reviewed to determine whether any new documents require manual
actions.
If a review is requested, a user assigned as a reviewer can access that taxonomy and add review
comments to the category. After the category is approved, either the editor or reviewer can request
publication.
The publisher opens the taxonomy and publishes the category. Every day, the publisher can access
the draft view to determine whether categories were updated and require republication. The publisher
can also use the published request filter to determine whether any categories require publication.
HPE IDOL Collaborative Classifier (11.1)
Page 143 of 155
User Guide
Appendix A: Workflow Example
HPE IDOL Collaborative Classifier (11.1)
Page 144 of 155
User Guide
Appendix B: HPE Category XML Format
Appendix B: HPE Category XML Format
This appendix describes the required structure of the XML file that you can use with HPE Category to
create categories.
•
•
•
Introduction
145
XML Format
145
Example Category XML Files
153
Introduction
There are two ways to use an XML file to create a category structure in HPE ICC:
l
Use the HPE IDOL Categorizer. To import a hierarchy from an XML file, run the
CategoryImportFromXML action. For example:
http://
IDOLhost
:port/action=CategoryImportFromXML&ImportFilename=MyCategory.xml&BuildNow=True
For information on importing category information by using the HPE Category component, see the
HPE IDOL Server Administration Guide.
l
Use HPE IDOL Collaborative Classifier. To import a hierarchy from an XML file, you must use the
Taxonomy module.
XML Format
This section lists and describes the tags that are allowed in the XML file from which you want to import
a category structure.
You must include the <autn:categories> tag in the XML file; however, there are no required tags
within <autn:categories>. If you include only the <autn:categories> tag, HPE ICC imports an
empty category structure. To expand the category structure, use the <autn:category> tag and its
children.
<autn:categories> (required)
The <autn:categories> tag marks the beginning of the XML categories that HPE IDOL Collaborative
Classifier reads. When you use the CategoryImportFromXML action, HPE IDOL Collaborative
Classifier reads the XML between the opening and closing <autn:categories> tags.
All categories (<autn:category>) are children of (<autn:categories>) when you export a category
from HPE ICC to XML. You can then import the XML to HPE ICC. They should have the same format.
You must include an XML namespace in the tag. For example:
<autn:categories xmlns:autn="http://schemas.autonomy.com/aci">
HPE IDOL Collaborative Classifier (11.1)
Page 145 of 155
User Guide
Appendix B: HPE Category XML Format
Tag name
Number allowed
Required
<autn:category>
one or more
No
<autn:category>
The <autn:category> tag marks the limits of each category that you want to import in your XML file.
You can include one or more <autn:category> tags inside the <autn:categories> tag, and
<autn:category> tags can also contain child <autn:category> tags.
The following table lists the tags that are allowed in <autn:category>.
Tag name
Number allowed
Required
<autn:name> (required)
one
Yes
<autn:id>
one
No
<autn:parent>
one
No
<autn:refersto>
one
No
<autn:trainingelement>
one or more
No
<autn:simplecat>
one
No
<autn:relevancecat>
one
No
<autn:details>
one
No
<autn:name> (required)
The <autn:name> tag sets the name of the category. You must include one <autn:name> within each
set of <autn:category> tags. For example:
<autn:name>UKpolitics</autn:name>
Tags allowed within <autn:name>: none
<autn:id>
The <autn:id> tag sets the HPE IDOL category ID. If a category with this ID already exists in the
server, HPE IDOL uses the OnConflict ACI parameter to determine the action to take.
For further information, see the CategoryImportFromXML OnConflict parameter in the HPE IDOL
Server Reference.
HPE IDOL Collaborative Classifier (11.1)
Page 146 of 155
User Guide
Appendix B: HPE Category XML Format
<autn:parent>
The <autn:parent> tag identifies the ID of the parent category. This option is effective only if you set
Flat to True in the CategoryImportFromXML action.
For further information, see the HPE IDOL Server Reference.
<autn:refersto>
The <autn:refersto> tag identifies the ID of the category to which the new category refers. This
information is used to create a category that refers to another, and which inherits its fields, training, and
special documents. This option is effective only if you set Flat to True in the CategoryImportFromXML
action.
For further information, see the HPE IDOL Server Reference.
<autn:trainingelement>
The <autn:trainingelement> tag identifies the training element for a category. HPE IDOL Server
identifies concepts that belong to the category from this training set. You can include one or more
<autn:trainingelement> tags in each set of <autn:category> tags.
The following table lists the tags that are allowed in <autn:trainingelement>.
Tag name
Number allowed
Required
<autn:type>
one
Yes
<autn:content>
one
See below
<autn:language>
one
No
<autn:reference>
one
See below
<autn:docid>
one
See below
<autn:database>
one
No
One or more of these:
<autn:type>
The <autn:type> tag sets the type of training to be used by <autn:trainingelement>. Each
<autn:trainingelement> must contain only one <autn:type> tag.
The following table describes the values that are valid for <autn:type>.
Options for
<autn:type>
Description
TRAININGTEXT
Identifies the training type as text only.
BOOLEAN
Identifies the training type as Boolean. You must define the Boolean operator in the
HPE IDOL Collaborative Classifier (11.1)
Page 147 of 155
User Guide
Appendix B: HPE Category XML Format
Options for
<autn:type>
Description
<autn:content> tag. For example:
<autn:type>BOOLEAN</autn:type>
<autn:content>(phone AND mobile)</autn:content>
URLDOWNLOAD
Identifies text from a URL to use for training.
DREDOCUMENT
Identifies the training type as a document indexed in HPE IDOL. The document is
specified by either autn:reference or autn:docid.
<autn:content>
The <autn:content> tag specifies the training text or Boolean expression. This tag is required if
<autn:type> is TRAININGTEXT, BOOLEAN, or URLDOWNLOAD.
For more information, see <autn:type>, on the previous page.
<autn:language>
The <autn:language> tag specifies the language of the training text. This tag is optional.
<autn:reference>
The <autn:reference> tag specifies the HPE IDOL DREREFERENCE to use for training. This tag is
required if <autn:type> is DREDOCUMENT, and you do not set <autn:docid>.
NOTE:
If the document is of the type DREDOCUMENT, you must set either <autn:reference> or
<autn:docid>, but not both.
For more information, see <autn:type>, on the previous page and <autn:docid>, below.
<autn:docid>
The <autn:docid> tag specifies the HPE IDOL DocID to use for training. This tag is required if
<autn:type> is DREDOCUMENT, and you do not set <autn:reference>.
NOTE:
If the document is of the type DREDOCUMENT, you must set either <autn:reference> or
<autn:docid>, but not both.
For more information, see <autn:type>, on the previous page and <autn:reference>, above.
<autn:database>
The <autn:database> tag specifies the HPE IDOL database in which the training document is located.
You can set only one database for each <autn:trainingelement> tag. This tag is optional.
<autn:simplecat>
The <autn:simplecat> tag specifies whether the category is a simple category. For example:
<autn:simplecat>True</autn:simplecat>
HPE IDOL Collaborative Classifier (11.1)
Page 148 of 155
User Guide
Appendix B: HPE Category XML Format
NOTE:
If you use this tag, do not set <autn:relevancecat>.
<autn:relevancecat>
The <autn:relevancecat> tag specifies whether the category is a relevance category. For example:
<autn:relevancecat>False</autn:relevancecat>
NOTE:
If you use this tag, do not set <autn:simplecat>.
<autn:details>
The <autn:details> tag sets training details for the category. You can include one set of
<autn:details> within each set of <autn:category> tags.
The following table lists the tags that are allowed in <autn:details>.
Tag name
Number allowed
Required
<autn:generatedterms> and <autn:generatedweights>
one
See below
<autn:queryagenttnw>
one
No
<autn:modifiedterms> and <autn:modifiedweights>
one
No
<autn:exclusions>
one
No
<autn:inclusions>
one
No
<autn:fakeweights>
one
No
<autn:numresults>
one
No
<autn:threshold>
one
No
<autn:databases>
one
No
<autn:fieldtext>
one
No
<autn:taxonomyroot>
one
No
<autn:active>
one
No
<autn:role>
one
No
<autn:memberpermissions>
one
No
<autn:nonmemberpermissions>
one
No
<autn:simplecatdefaultcat>
one
No
<autn:relevantcat>
one
No
HPE IDOL Collaborative Classifier (11.1)
Page 149 of 155
User Guide
Appendix B: HPE Category XML Format
Tag name
Number allowed
Required
<autn:simplecatparam>
one
No
<autn:userfields>
one
No
<autn:generatedterms>
The <autn:generatedterms> tag sets terms generated for a category from training. Use this feature
only if you are editing an existing category from which you can take the terms. You can include one
<autn:generatedterms> in each set of <autn:details> tags. For example:
<autn:generatedterms>LYMPH,MISDIAGNOS,PATHOLOGI</autn:generatedterms>
NOTE:
If you specify terms for a category in the <autn:generatedterms> tag, you must enter a
corresponding list of weights by using the <autn:generatedweights> tag.
The <autn:generatedterms> tag is required if you set <autn:queryagenttnw>. For more information,
see <autn:queryagenttnw>, below.
<autn:generatedweights>
The <autn:generatedweights> tag sets weights generated for the terms for the category from
training. Use this feature only if you are editing an existing category from which you can take the
weights. You can include one <autn:generatedweights> in each set of <autn:details> tags. For
example:
<autn:generatedweights>5960,4035,4001</autn:generatedweights>
NOTE:
If you specify weights for a category in the <autn:generatedweights> tag, you must enter a
corresponding list of terms by using the <autn:generatedterms> tag.
The <autn:generatedweights> tag is required if you set <autn:queryagenttnw>. For more
information, see <autn:queryagenttnw>, below.
<autn:queryagenttnw>
The <autn:queryagenttnw> tag specifies the query string generated by terms and weights used to
build the category. For example:
<autn:queryagenttnw>LYMPH~[5960] MISDIAGNOS~[4305] PATHOLOGI~[4001]
</autn:queryagenttnw>
NOTE:
You can use <autn:queryagenttnw> with either <autn:generatedterms> and
<autn:generatedweights>, or <autn:modifiedterms> and <autn:modifiedweights>. If
both sets exist, modified terms take precedence over generated ones.
The <autn:queryagenttnw> tag performs the same function as the HPE IDOL CategoryBuild action.
For more information, see the HPE IDOL Server Administration Guide.
For more information, see <autn:generatedterms>, above, <autn:generatedweights>, above,
<autn:modifiedterms>, on the next page, and <autn:modifiedweights>, on the next page.
HPE IDOL Collaborative Classifier (11.1)
Page 150 of 155
User Guide
Appendix B: HPE Category XML Format
<autn:modifiedterms>
The <autn:modifiedterms> tag sets terms defined by the user. You can include one
<autn:modifiedterms> within each set of <autn:details> tags. For example:
<autn:modifiedterms>LYMPH,MISDIAGNOS,PATHOLOGI</autn:modifiedterms>
For more information on changing the weights of terms in a category, see the HPE IDOL Server
Administration Guide.
NOTE:
If you specify terms for a category in the <autn:modifiedterms> tag, you must enter a
corresponding list of weights by using the <autn:modifiedweights> tag.
<autn:modifiedweights>
The <autn:modifiedweights> tag sets weights for terms defined by the user. You can include one
<autn:modifiedweights> within each set of <autn:details> tags. For example:
<autn:modifiedweights>5960,4035,4001</autn:modifiedweights>
For more information on changing the weights of terms in a category, see the HPE IDOL Server
Administration Guide.
NOTE:
If you specify weights for a category in the <autn:modifiedweights> tag, you must enter a
corresponding list of terms by using the <autn:modifiedterms> tag.
<autn:exclusions>
A comma-separated list of documents to exclude from category queries. For example:
<autn:exclusions>C:\temp\doc1.txt,C:\temp\doc2.txt</autn:exclusions>
For more information, see the CategorySetSpecialDocs Exclusions parameter in the HPE IDOL
Server Reference.
<autn:inclusions>
The <autn:inclusions> tag contains a comma-separated list of documents to include in category
queries. For example:
<autn:inclusions>C:\temp\docA.txt,C:\temp\docB.txt</autn:inclusions>
For more information, see the CategorySetSpecialDocs Inclusions parameter in the HPE IDOL
Server Reference.
NOTE:
If you use this tag, you must set corresponding weights for the included terms by using the
<autn:fakeweights> tag.
<autn:fakeweights>
The <autn:fakeweights> tag contains a comma-separated list of weights for documents specified in
<autn:inclusions>. Weights must correspond to the documents in number and order. For example:
<autn:fakeweights>800,2200</autn:fakeweights>
HPE IDOL Collaborative Classifier (11.1)
Page 151 of 155
User Guide
Appendix B: HPE Category XML Format
For more information, see the CategorySetSpecialDocs Fakeweights parameter in the HPE IDOL
Server Reference.
<autn:numresults>
The <autn:numresults> tag sets the maximum number of results that category queries can return.
You can include one <autn:numresults> tag for each category within <autn:details> tags. For
example:
<autn:numresults>10</autn:numresults>
<autn:threshold>
The <autn:threshold> tag sets the minimum relevance score that documents must possess to
appear in category query results. You can include one <autn:threshold> tag for each category within
<autn:details> tags. For example:
<autn:threshold>400</autn:threshold>
<autn:databases>
The <autn:databases> tag sets the databases in which documents must exist to appear in category
query results. You must separate multiple databases with plus symbols, commas, or spaces. For
example:
<autn:databases>Archive,Minicar</autn:databases>
<autn:fieldtext>
The <autn:fieldtext> tag specifies the fields that result documents must contain, and the conditions
that these fields have to meet for the documents to be returned as results.
For more information, see the CategorySetFields Fieldtext parameter in the HPE IDOL Server
Reference.
<autn:taxonomyroot>
The <autn:taxonomyroot> tag specifies whether the category is a taxonomy root. For example:
<autn:taxonomyroot>True</autn:taxonomyroot>
<autn:active>
The <autn:active> tag specifies whether the category is active. For example:
<autn:active>True</autn:active>
<autn:role>
The <autn:role> tag specifies the role or roles that you want to give access to the category. For
example:
<autn:role>Usertype1,Usertype2</autn:role>
Use <autn:memberpermissions> and <autn:nonmemberpermissions> to set the category access
permissions for members and non-members of this role.
<autn:memberpermissions>
The <autn:memberpermissions> tag sets the category access permissions that you want role
members to have. Separate multiple permissions with commas. For example:
<autn:memberpermissions>read,edit</autn:memberpermissions>
HPE IDOL Collaborative Classifier (11.1)
Page 152 of 155
User Guide
Appendix B: HPE Category XML Format
For more information, see the CategorySetPermissions MemberPermissions parameter in the HPE
IDOL Server Reference.
<autn:nonmemberpermissions>
The <autn:nonmemberpermissions> tag sets the category access permissions that you want role
non-members to have. Separate multiple permissions with commas. For example:
<autn:nonmemberpermissions>read</autn:nonmemberpermissions>
For more information, see the NonMemberPermissions parameter for the CategorySetPermissions
action in the HPE IDOL Server Reference.
<autn:simplecatdefaultcat>
The <autn:simplecatdefaultcat> tag specifies which of the category's children is to be the default
category for CategorySimpleCategorize. You must use the category ID value. For example:
<autn:simplecatdefaultcat>1230982349874568</autn:simplecatdefaultcat>
For more information, see the CategorySetDetails SimpleCatDefaultCat parameter in the HPE
IDOL Server Reference.
<autn:relevantcat>
The <autn:relevantcat> tag specifies which of the category's children is to be used as the relevant
category. You must use the category ID value. For example:
<autn:relevantcat>324987602</autn:relevantcat>
For more information, see the CategorySetDetails RelevantCat parameter in the HPE IDOL Server
Reference.
<autn:simplecatparam>
The <autn:simplecatparam> tag sets a numeric factor that increases or decreases the probability of
this category being chosen by the CategorySimpleCategorize action. For example:
<autn:simplecatparam>1.4</autn:simplecatparam>
For more information, see the CategorySetDetails SimpleCatParam parameter in the HPE IDOL
Server Reference.
<autn:userfields>
The <autn:userfields> tag sets fields and values defined by the user. For example:
<autn:userfields>
<autn:acc-inbox-threshold>30</autn:acc-inbox-threshold>
</autn:userfields>
Example Category XML Files
The following is an example XML file that includes only required elements:
<?xml version="1.0" encoding="UTF-8" ?>
<autn:categories xmlns:autn="http://schemas.autonomy.com/aci/">
<autn:category>
HPE IDOL Collaborative Classifier (11.1)
Page 153 of 155
User Guide
Appendix B: HPE Category XML Format
<autn:name>MyCategory</autn:name>
</autn:category>
</autn:categories>
The following is an example of a Category XML file that includes one category:
<?xml version="1.0" encoding="UTF-8" ?>
<autn:categories xmlns:autn="http://schemas.autonomy.com/aci/">
<autn:category>
<autn:name>test</autn:name>
<autn:trainingelement>
<autn:type>TRAININGTEXT</autn:type>
<autn:content>hotels paris food stay</autn:content>
<autn:language>ENGLISH</autn:language>
</autn:trainingelement>
<autn:details>
<autn:generatedterms>STAI,PARI,FOOD,HOTEL</autn:generatedterms>
<autn:generatedweights>2352,2272,1884,417</autn:generatedweights>
<autn:queryagenttnw>FOOD~[1884] HOTEL~[417] PARI~[2272]STAI~[2352]
</autn:queryagenttnw>
<autn:active>True</autn:active>
</autn:details>
</autn:category>
</autn:categories>
HPE IDOL Collaborative Classifier (11.1)
Page 154 of 155
Send documentation feedback
If you have comments about this document, you can contact the documentation team by email. If an email
client is configured on this system, click the link above and an email window opens with the following
information in the subject line:
Feedback on User Guide (HPE IDOL Collaborative Classifier 11.1)
Add your feedback to the email and click Send.
If no email client is available, copy the information above to a new message in a web mail client, and send
your feedback to [email protected]
We appreciate your feedback!
HPE IDOL Collaborative Classifier (11.1)
Page 155 of 155
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