Taxonomies And Ontologies

1
Taxonomies And Ontologies

• An Exploratory Overview

T.B. Rajashekar National Centre for Science
InformationIndian Institute of ScienceBangalore
560 012 (raja_at_ncsi.iisc.ernet.in)
Prepared for the talk at the Sarada Ranganathan
Endowment for Library ScienceBangalore 9th
August 2003
2
Agenda

• To highlight
• Growing importance of taxonomies and ontologies
• Definitions, methodologies, tools, examples,
  experiences, Trends
• Draw implications for information professionals

3
Growing Importance

• Google search
• Taxonomy/ies 186,000 pages
• Ontology/ies 202,000 pages
• Library Classification 1,160,000
• Thesauri/us 899,000
• WOS search shows significant increase in number
  of publications related to taxonomies and
  ontologies during 2000 and 2001

4
Why this Interest?

• Rapidly growing volume of digital information
• Within enterprises
• Internet
• Enterprises
• Effective management and access to intellectual
  capital of the enterprise (corporate memory)
• Internet
• Access and exchange of meaningful information and
  data human beings, between machines and
  services (software agents)
• Semantic web - Content schemas metadata,
  encoding schemes, vocabulary

5
Definitions - Taxonomy

• Definitions vary widely in scope from simple
  and narrow to complex and broad
• Roots in biological taxonomy
• Adapted mainly by KM, enterprise portal, content
  management and CS community
• Beginning to be used and accepted in information
  science
• Lets peruse a few definitions given by KM and
  taxonomy experts

6
Definitions - Taxonomy

• Systematic way of classifying knowledge
• Provides a hierarchical structure of concepts,
• using terms that help in the development of a
  common language to aid knowledge sharing
• - Taxonomies Framework for Corporate Knowledge,
  by Jan Wyllie (2003)

7
Definitions Taxonomy

• A correlation of different functional languages
  used by the enterprise
• to support a mechanism for navigation and gaining
  access to the intellectual capital of the
  enterprise
• by providing such tools as portal navigation
  aids, authority for tagging documents and other
  information objects,
• support for search engines, and knowledge maps
• and possibly, a knowledge base in its own right.
• Enterprise Taxonomy
• - Alan Gilchrist and Peter Kibby, TFPL

8
Definitions Taxonomy

• Structures that provide a way of classifying
  things living organisms, products, books
• into a series of hierarchical groups to make them
  easier to identify, study, or locate.
• Taxonomies consist of two parts structures and
  applications.
• Structures consist of the categories (or terms)
  themselves and the relationships that link them
  together.
• Applications are the navigation tools available
  to help users find information
• - Jean Graef, Montague Institute

9
Definitions Taxonomy

• The specification of the names of people, places,
  things, and anything else that is needed to allow
  search engines and other content applications to
  work better.
• - Joseph A. Busch, Taxonomy Strategies.

10
Definitions Taxonomy
Taxonomy Facets
- Joseph A. Busch, Taxonomy Strategies
11
Definitions Taxonomy

• Taxonomy a classification of elements within a
  domain (read corporate domain)
• Domain a sphere of knowledge, influence, or
  activity
• Classification the operation of grouping
  elements and establishing relationships between
  them (or the product of that operation)
• Relationships a defined linkage between two
  elements
• Element an object or concept

- Mike Crandall, Microsoft Information Services
12
Definitions Taxonomy

• Global Knowledge Object (GKO) taxonomy standard
  (standard data architecture and vocabulary ).
• PwCs trade name for their standard
  unstructured data architecture
• Establishes a defined series of fields
• Defines vocabularies allowing consistent tagging
  of content based on business usage and context
• Includes multiple taxonomies covering
  unstructured data/content
• Acts as a logical data model representing what
  we need to know about our content
• Provides a standard approach for meta-data
  tagging of content to enable repurposing in
  multiple systems
• GKO will support multi-term and multi-language
  vocabulary sets, allowing differences across
  business units and languages

- Mark Zoeckler, Director, PricewaterhouseCoopers
13
Definitions Taxonomy

• Taxonomies are evolving to be much more than
  traditional classification systems
• Support structure, content and applications
  (navigation tools)
• Customized to reflect the language, culture and
  goals of particular enterprise
• Often created using a combination of human
  efforts and software
• Reflect disparate information resources
  e-mails, memos, people, documents, etc.
• Part of a process constantly refined and updated

14
Definitions - Ontology

• Ontologies are more complex than taxonomies, and
  express relationships between the elements of a
  taxonomy, such as part of or located in.
• Theory
• Ontology Science or study of being
• Term used to refer to the shared understanding of
  some domain of interest, as a set of concepts
  (e.g. entities, attributes, processes), their
  definitions and their interrelationships world
  view (conceptual ontology)

15
Definitions Ontology

• An explicit ontology typically includes a
  vocabulary of terms and some specification of
  their meaning (i.e. definitions) and
  relationships (Rokhlenko Oleg, Data Integration
  Seminar, Spring 2002)
• Why ontologies?
• Communication between people and organizations
  within a domain
• Interoperability between systems
• Prerequisite for the semantic web

16
The term procedure used by one tool is
translated into the term method used by the
other via the ontology, whose term for the same
underlying concept is process.
Example (From Rokhlenko Oleg, Data Integration
Seminar, Spring 2002)
procedure viewer
give me the procedure for
give me the process for
translator
procedure ???
here is the procedure for
Ontology
??? process
give me the METHOD for
procedure process
translator
METHOD process
here is the process for
method library
here is the METHOD for
17
Building (corporate) Taxonomies
18
Building Taxonomies - 1
Mike Crandall, Microsoft Information Services

• Define project scope
• Define boundaries, determine required resources
• Obtain resource commitments
• User needs survey
• What content users need, how they access it
• Information audit
• Existing content, its structure, who is
  responsible
• Involve users
• Include key stakeholders in the process
• Decide on architectural approach how you are
  going to do it
• Build taxonomy

19
Building Taxonomies - 1
Mike Crandall, Microsoft Information Services
The Process

• Identify
• business
• needs
• _______
• User
• needs
• survey
• Tag
• audit
• Content
• audit

• Collect/
• structure
• terms
• ________
• Build
• vocabs
• Define
• rules
• Create
• change
• control
• process

• Tag
• content
• ________
• Embed
• vocab
• access
• in tools
• Provide
• guidelines
• for use

• Expose
• Content
• ________
• Embed
• tags in
• interfaces
• Segment
• content by
• attributes
• Enable
• thru
• XML/XSL

• Define
• needed
• attributes
• _______
• Build
• object
• model
• Create
• flat list
• Provide
• mapping
• schema?

20
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies

• Phase 1 Scope taxonomy
• Brief, identify and interview stakeholders and
  subject matter experts.
• Collect example documents.
• Discover existing controlled vocabularies.
• Analyze data (inductive and deductive).
• Present scoping results.
• Phase 2 Build taxonomy
• Develop broad taxonomy outline (1-3 levels deep)
• Review, revise, and approve taxonomy outline with
  stakeholders and subject matter experts.
• Fill in taxonomy outline (with approx 1500 terms)
  
• Review, revise, and approve draft taxonomy with
  stakeholders and subject matter experts.

21
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies
Taxonomy outline example (MS Word)
22
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies

• Phase 3 Maintain taxonomy
• Develop tagging rules and procedures.
• Specify taxonomy maintenance business process.
• Document taxonomy maintenance procedures.
• Train users.

23
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies
Taxonomy maintenance Editorial rules
24
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies
Taxonomy maintenance -business process
25
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies
Taxonomy maintenance change scenario
old old label to be deleted new new label
to take its place Remember to match using the
full path of the code, since parts of codes may
have similar names. for each file in the
staging area V vpath for the file
system(iwmetadata -get V) gt tmpFile
updateFlag false if (tmpFile contains
old) Note that if you are renaming an
internal term, a file may have multiple
concepts with that internal label, so be
sure to loop through them all.
s/old/new/ updateFlag true
if (updateFlag true) system(iwmetadata
set tmpFile V) Listing 1 Simple
metadata update pseudocode.
26
Building Taxonomies - 2
Joseph A. Busch, Taxonomy Strategies
Taxonomy Example
27
Use of Taxonomies 1 Microsoft Information
Services
28
Use of Taxonomies 1
Microsoft Information Services

• Content creation- tagging of documents and other
  items
• Site navigation- using categories
• Information retrieval - search
• Personalization- delivery
• Link content to people via user profiles
  requires well tagged content

29
Content Tagging
30
Tagged Content
31
Category-based Navigation
32
(No Transcript)
33
Categories in Search
34
(No Transcript)
35
Use of Taxonomies 2 Joseph A. Busch Taxonomy
Strategies
36
Case Study Halliburton

• Consolidate multiple taxonomies that have been
  developed for separate product lines so that
  employees, contractors and customers can easily
  find information about products and services
  regardless of what product line theyre in.
• Find everything on a piece of equipment before
  performing a service.
• Find a specific piece of information that you
  know is there (or should be there).
• Help communities of practice organize detailed
  technical information in their intranet portals.
• Help organize and present search results in
  categories and sub-categories

37
Halliburton Taxonomy

• Key data drivers are Business Process and
  Logistics (materials and equipment).

Case Study
Logistics
38
New Halliburton facet search browser
39
Case Study NASA

• Make it easy for various audiences to find all
  the relevant information from all the NASA
  programs quickly.
• Provide one stop shopping for NASA resources.
• Provide search results targeted to user
  interests.
• Help users to easily find links to databases and
  tools.
• Inspire the next generation of explorers.
• Prepare Americas future in innovation,
  technology and exploration.
• Advance e-government strategies through projects
  that are citizen-centric.

40
NASA Taxonomy

• Key data drivers are business roles, functions,
  and skills.

Case Study
41
Use of Taxonomies 3 TopicMap in Highwire
E-Journal Service Automatic tagging Semio
Tagger (Entrieva) Visualization Using
Hyperbolic Tree technology of Inxight Software
Visualization
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
Benefits of Taxonomies

• Improved management of unstructured information
• Provide context (additional value)
• Simplify navigation
• Improve search
• Enable personalization
• Share information between systems and people in
  B2B e-commerce
• Make connections between related concepts across
  disciplines/ applications/ services

47
Taxonomy Software

• Approaches for taxonomy generation
• Training by example Editor defines a category
  is and provides a training set
• Rules-based documents are classified according
  to specified business rules
• Statistical methods word patterns are
  identified according to parameters such as
  frequency of use and proximity to other words and
  phrases (clustering)
• Natural language processing (computational
  linguistics) classification relies on extensive
  dictionary and thesaurus for identifying concepts

48
Taxonomy Software

• Key features supported by taxonomy software
• Provide a variety of user control and interaction
• Suggest a classification that the user can accept
  or override
• Co-exist with content-management and
  search-and-retrieval software
• Most taxonomy tools automatically insert metadata
  as XML tags into documents
• Visualization graphical presentation of taxonomy

49
Taxonomy Software

• Vendors
• Heavy weights players whose product suites
  encompass whole lifecycle of information,
  including taxonomy creation, document
  classification and information retrieval (Verity,
  Autonomy)
• Specialist taxonomy suppliers using innovative
  approaches
• Consolidation in the market place between
  classification tool vendors and search engine
  providers

50
Standards

• Interoperability
• Need for interchange of information between
  organizations in a commonly agreed way
• XML - basic level of information exchange
  standard
• Need for broad industry-wide consortia to use
  agreed schemas for sharing data across
  applications
• Related standards
• RDF framework for describing and using metadata
  schemes (e.g. Dublin Core)
• OIL (Ontology Inference Layer)
• WSDL (Web Services Definition Language), UDDI
  (Universal Description, Discovery and
  Integration), ISO Topic Maps

51
Taxonomy Development Some Lessons
What experts are saying?

• Precise methodologies do not exist still an art
  form
• Invest in small, pilot projects stakeholders
  can judge success, potential for quick success
• Use people with domain knowledge and library
  skills
• Keep it simple a good taxonomy is intuitive and
  intelligible

52
Taxonomy Development Some Lessons
What experts are saying?

• Involve users at all aspects of development and
  usage lifecycle
• Focus on fitness for purpose better to have
  several application/user specific taxonomies than
  one-size-fits-all approach
• Must be constantly updated based on usage and
  changing terminology
• Use technology to identify patterns and
  terminology, but use human intervention for final
  judgment

53
Taxonomy Development Some Challenges
What experts are saying?

• Overall number of nodes
• Right balance between breadth and depth
• Choosing appropriate hierarchical structures
• To start from scratch or use an off-the-shelf
  taxonomy
• Choose between automation and specialist human
  classifiers

54
Future Directions
What experts are saying?

• Taxonomies and ontologies are closely linked to
  the future of the web - widely accepted
  taxonomies of hundreds (if not thousands) of
  different knowledge domains are the building
  blocks of the future semantic web.
• Intelligent web will incorporate topic maps,
  knowledge maps and ontologies that act on the
  basis of the precise meaning of specified terms
  and the relationship between them.
• It is also expected that, in addition to an
  intelligence being artificially situated in the
  network, it also has to come from enhancing the
  intelligence, disciplines and skills of the users
  using taxonomies and search engines.

55
Implications for Information Professionals

• Huge demand for taxonomy specialists, portal/
  intranet/ website content management, metadata,
  vocabulary development
• How can we contribute?
• Enhance (corporate/ web) taxonomy, ontology
  development theory, principles and procedures
  with those of classification and thesauri design
• Evolve and impart to our students competencies in
  developing enterprise taxonomies, metadata
  schemas and integration of these into enterprise
  IKM systems (portals, CMS, DMS, DLs, etc.)

56
Implications for Information Professionals

• Lessons for us?
• Systematization of taxonomy development and
  maintenance
• Dynamic nature of enterprise taxonomies
• Awareness of existing taxonomies and vocabularies
  and their adoption (e.g. taxonomywarehouse.com)

57
Thank You!