Академический Документы
Профессиональный Документы
Культура Документы
FROM THESAURUS TO ONTOLOGY
TO SEMANTIC WEB:
ONTOLOGIES AS ESSENTIAL
BACKBONE TO SEMANTIC SEARCH.
Juhana Salim
Faculty of Information Science &
Faculty of Information Science &
Technology
Universiti Kebangsaan Malaysia
1
2/22/2011
Why thesauri
Conceptual and vocabulary problems users faced
when searching the web
Queries illustrating these problems
Synonym expansion and Hierarchic expansion
• Query 1. Drug use by teenagers
• Query 1.1 teenage* AND drug*
• Query 1.2. Synonym expansion for teenage
teenage*
• (teenage* OR teen OR teens OR youth* OR
• adolescent* OR kid* OR "high school")
• AND drug*
Query 1.1. teenage* AND drug*
(AltaVista)
‐ ‐.
About 30 documents match your query.
1. CEIDA Druglinks ‐ Info Centre ‐ PARENTS TALKING TO TEENAGERS ABOUT
DRUGS
What do parents want from their teenagers? Basically, parents want: To know your
kids
are alright and not in danger. To know your kids think you're OK...
http://www. ceida. net. au/info_centre/drug~myths/what_do. html ‐ size 3K ‐ 21‐May‐
97 ‐
English
2. CEIDA Druglinks ‐ Info Centre ‐ PARENTS TALKING TO TEENAGERS ABOUT
DRUGS
Better Ways of Communicating. Different points of view Communication is the key to
Better Ways of Communicating Different points of view Communication is the key to
resolving problems, if they exist. Or to finding out if they exist....
http1A~www. ceida. net. au/info_centre/drug~myths/better.html ‐ size 9K ‐ 21‐May‐97
‐
English
2
2/22/2011
Query 1.2. Synonym expansion of teenager
( teenage* OR teen OR teens OR youth OR adolescent* OR kid* OR "high
school")AND drug *
About 249 documents match your query.
1 Ad l D Ab T O
1. Adolescent Drug Abuse Treatment Outcome
Adolescent Drug Abuse Treatment Outcome. Executive Summary. This is a
report on the
evaluation of an inpatient adolescent drug abuse treatment program in..
http://www. cbc. med. umn. edu/~andy/drugabuse/adoltx. htm ‐ size 3K ‐ 28‐
Sep‐96 ‐
English
2. Poll finds parents overestimate communication with kids on drugs
03/03/97 ‐ 07:26 PM ET ‐ Click reload often for latest version. Poll finds
parents
overestimate communication with kids on drugs. NEW YORK ‐ Most parents..
http://cgi.usatoday.com/elect/eq/eq17&htm ‐ size 2K ‐ 21‐May‐97 ‐ English
Query 1.3. Plus synonym and hierarchic expansion of “drug*”
( teenage* OR teen OR teens OR youth* OR adolescent* OR kid*
OR "high school")
AND ( drug* OR substance* OR alcohol OR nicotine OR smoking
OR cigarette*)
About 409 documents match your query.
About 409 documents match your query.
1. Smoking is NOT for kids!
We believe smoking is for adults only. We therefore require that you
be at least 18 years of
age in order to view this site. Click below to enter the...
http://www.smokers.org/ ‐ size 820 bytes ‐ 20‐Apr‐97 ‐ English
2. Adolescent Drug Abuse Treatment Outcome
Adolescent Drug Abuse Treatment Outcome. Executive Summary. This
Adolescent Drug Abuse Treatment Outcome. Executive Summary. This
is a report on the
evaluation of an inpatient adolescent drug abuse treatment program
in..
http://www cbc. med. umn. edu/~andy/drugabuse/adoltx. htm ‐ size
3K ‐ 28‐Sep‐96 ‐
English
3
2/22/2011
Query 2: classification
Examples from Lycos search
139) RESIDENCE CLASSIFICATION
139) RESIDENCE CLASSIFICATION
Residence Classification Residence Classification Nonresident students seeking to
become California residents for tuition/fee purposes must petition t.
http://www.reg.uci.edu/REGISTRAR/SOC/rc.html [99%]
152) PRODUCT CLASSIFICATION
EPA may classify a pesticide product for restricted use if its characteristics warrant
special handling. Restricted use pestici.
http://hammock.ifas.ufl.edu/txt/fairs/26668 [99%]
http://hammock.ifas.ufl.edu/txt/fairs/26668 [99%]
426) Dewey Decimal Classification Home Page
DDC 21 and Dewey for Windows now available! OCLC Forest Press is pleased to
announce the publication of DDC 21, the latest edition of the Dewey Decima.
http://www.oclc.org/fp/ [99%]
Query 3.1. classification and security
Examples from AltaVista search
Restricts results but also misses a lot.
1. EXSYS: Specific Applications: Security Classification
Nuclear Weapons Security Classification. US Dept. of Energy. Nuclear...
http://www.exsysinfo.com/Appnotes/nuclear.html ‐ size 7K ‐ 22‐May‐97 – English
2. SLATE Application Note ‐‐Security Classification and Automatic Page Marking wi
Introduction. If your document contains classified information, you can identify the
classification by.
http://www.slate.tdtech.com/app_notes/secclass‐html.html ‐ size 6K ‐ 22‐Feb‐96 ‐
p y
3. Computer Security Classification
The Classification. alert Advisories on various security vulnerabilities. dict Dictionaries
and word lists. doc Security related documents.
access_control.
http://www.cs.purdue.edu/coast/archive/Classification
4
2/22/2011
How can we assist user with search
topic clarification?
Thesaurus can be used as knowledge base for
an interface that can assist user with search
topic clarification
Example from the Art and
Architecture Thesaurus
• <art genres>
• academic art
• amateur art
• apocalyptic art
• art brut
• children's art
• commercial art
• community art
• SN Includes art undertaken in with particular communities, often socially
deprived, usually with the idea of producing an effect or inspiring
response specifically within those communities, with no reference to
widely established standards.
5
2/22/2011
Example from the Art and
Architecture Thesaurus
• For art intended to beautify or
• enrich public places, use public
enrich public places, use public
• art.
• computer art
• court art
• crafts
• cybernetic art
• did i
didactic art
• dissident art
• ethnic art
Sample documents with descriptors
Document
• The drug was injected into the aorta
• User concept: Systemic administration
U S i d i i i
Document:
• The percentage of children of blue‐collar workers
• going to college
• User concept: Intergenerational social mobility
Document:
• CSF studies on alcoholism and related behaviors
• User concept: Biochemical basis of behaviour
User concept: iochemical basis of behaviour
• User concept: longitudinal study
(Longitudinal not mentioned in the document;
determined through careful examination of the
methods section.)
6
2/22/2011
What is a thesaurus
A thesaurus is a structure that
manages the complexities of terminology and
• manages the complexities of terminology and
• provides conceptual relationships,
• ideally through an embedded
classification/ontology.
A thesaurus may specify descriptors authorized for
• indexing and searching. These descriptors form a
• controlled vocabulary (authority list, index
language)
Examples of classifications and thesauri
Alcohol and Other Drug Thesaurus (AOD
Thesaurus)
(US Nat. Inst. of Alcohol Abuse and Alcoholism)
http://etoh.niaaa.nih.gov/AODVol1/Aodthome.htm
Medical Subject Headings (MeSH) and
Unified Medical Language System (UMLS)
(US National Library of Medicine)
www.nlm.nih.gov/mesh/meshhome.html, www.nlm.nih.gov/mesh/MBrowser.html,
www.nlm.nih.gov/research/umls/umlsmain.html, http://umlsinfo.nlm.nih.gov
Art and Architecture Thesaurus (AAT)
(Getty Foundation)
http://www.getty.edu/research/tools/vocabulary/aat/index.html
7
2/22/2011
More examples of classifications and thesauri
Dewey Decimal Classification
(US Library of Congress and OCLC/Forest Press)
(US Library of Congress and OCLC/Forest Press)
http://www.oclc.org/dewey/about/ddc_21_summaries.h
tm
WordNet (Princeton University, George Miller)
www.cogsci.princeton.edu/~wn/,
www.notredame.ac.jp/cgi‐bin/wn (Not reachable on July
6, 2002)
6 2002)
CYC Ontology (CYC Corporation)
http://www.cyc.com/cyc‐2‐1/cover.html,
http://www.cyc.com/cyc‐2‐1/toc.html
Additional examples illustrating
different functions
HS Harmonized Commodity Description and Coding System. World Customs
Organization, Brussels. Info: http://pacific.commerce.ubc.ca/trade/HS.html
NAICS North American Industrial Classification System
NAICS North American Industrial Classification System
"common industry definitions for Canada, Mexico, and the US. Developed in cooperation
with the US Economic Classification Policy Committee, Statistics Canada, and Mexico's
Info: www.census.gov/epcd/www/naics.html, www.naics.com
ICD‐10 The International Statistical Classification of Diseases and Related Health
Problems, tenth revision. Produced by the World Health Organization. Published in
many languages. Info: www.who.int/whosis/icd10/index.html,
www.cdc.gov/nchs/about/major/dvs/icd10des.htm
CPT Physicians' Current Procedural Terminology. CPT 2003. American Medical
Association. November 2002
Association November 2002
(Info: http://www.ama‐assn.org/ama/pub/category/3113.html,
listing of codes https://webstore.ama‐assn.org/index.jhtml)
8
2/22/2011
Functions of a thesaurus / classification /
ontological knowledge base
• Support learning and assimilating information.
• Assist researchers and practitioners with problem
Assist researchers and practitioners with problem
clarification.
• Support information retrieval.
• Provide knowledge‐based support for end‐user
searching.
• Support meaningful information display.
• Provide a tool for indexing.
Provide a tool for indexing
• Facilitate the combination of multiple databases or
unified access to multiple databases.
• Support document processing after retrieval.
Provide classification for action
• This list addresses the functions of formal
classifications In a broader perspective
classifications. In a broader perspective,
• classification is the basis for much of everyday
action, where we put people, things, and
• events in certain categories and, based on these
categories, predict the behavior of persons
• and things and the course and effects of events,
and things and the course and effects of events
determine our attitudes towards them, and
• plan action accordingly.
9
2/22/2011
Examples
• a classification of diseases for diagnosis,
• aa classification of medical procedures for
classification of medical procedures for
insurance billing,
• a classification of medical outcomes to assist with
treatment evaluation,
• a classification of commodities for customs,
• a classification of educational objectives for
l ifi ti f d ti l bj ti f
instructional development, Provide classification
for action
A classification of diseases for diagnosis
10
2/22/2011
A classification of medical procedures for
insurance billing
Aging ‐ Refers to the unpaid insurance claims or patient balances that are due past 30 days.
Most medical billing software's have the ability to generate a separate report for insurance
aging and patient aging. These reports typically list balances by 30, 60, 90, and 120 day
increments.
Appeal ‐ When an insurance plan does not pay for treatment, an appeal (either by the
provider or patient) is the process of formally objecting this judgment. The insurer may
require additional documentation.
Applied to Deductible ‐ Typically seen on the patient statement. This is the amount of the
charges, determined by the patients insurance plan, the patient owes the provider. Many
plans have a maximum annual deductible that once met is then covered by the insurance
provider.
Assignment of Benefits ‐ Insurance payments that are paid to the doctor or hospital for a
patients treatment
patients treatment.
Beneficiary ‐ Person or persons covered by the health insurance plan.
Clearinghouse ‐ This is a service that transmits claims to insurance carriers. Prior to
submitting claims the clearinghouse scrubs claims and checks for errors. This minimizes the
amount of rejected claims as most errors can be easily corrected. Clearinghouses
electronically transmit claim information that is compliant with the strict HIPPA standards
(this is one of the medical billing terms we see a lot more of lately).
Source: http://ezinearticles.com/?Medical‐Billing‐Terms‐and‐Medical‐Coding‐
Terminology&id=4974223
A classification of medical outcomes to
assist with treatment evaluation
11
2/22/2011
A classification of commodities for customs
A classification of educational objectives
for instructional development
Bloom’s Taxanomy
12
2/22/2011
Is it only Thesaurus?
However, if we want to create a knowledge‐
rich description of for example an (image of
i hd i ti ff l (i f
an) art object, medical, business, crime etc.
such as required by the "semantic web",
thesauri turn out to provide only part of the
knowledge needed.
Symbiosis of Thesaurus and Ontology
Both have been working toward the same set of goals.
New
Naming Semantic Web
concepts Communities
> 100
Naming years
entities Library Communities
13
2/22/2011
Semantic Web Technologies
Library
Communities “I have a dream for the Web [in
which computers] become
capable of analyzing all the
bl f l ll h
data on the Web – the content,
links, and transactions between
people and computers. A
Semantic Web, which should
make this possible, has yet to
emerge, but when it does, the
day‐to‐day
day to day mechanisms of
mechanisms of
trade, bureaucracy and our
daily lives will be handled by
machines talking to machines.
The intelligent agents’ people
have touted for ages will finally
materialize.”
Berners‐Lee’s vision
Semantic Web Technologies
Each layer is
dependent on the
layer below it
Figure 1 Semantic Web Stack
Development on the various levels has been a long time in the making
14
2/22/2011
Why the slow rate of development?
Music Ontology
Gene Ontology
Initiatives
Web’ fying Thesaurus
Converting existing tools Semantic web standards
<Ontology
support ontologyIRI="http://example.com/te
a.owl" ...> <Prefix name="owl"
IRI="http://www.w3.org/2002/07/ow
l#"/>
l# /
<Declaration>
<Class IRI="Tea"/>
</Declaration>
</Ontology>
Limitless potentials
Outside scope of W3Cs concern
15
2/22/2011
Subject & Genre Vocabularies for the
Semantic Web
1. Alcohol and Other Drug Thesaurus
(AOD Thesaurus: US Nat. Inst. Of
Alcohol Abuse and Alcoholism
2. Medical Subject Heading (MeSH) and
These vocabularies
Unified Medical Language System
present tremendous
(UMLS): US National Library of
potential
Medicine
• Improve access to web
3. Art and Architecture Thesaurus (AAT):
resources and Semantic
Getty Foundation
web data.
4. Dewey Decimal Classification : US
• Enhanced network
Lib
Library of Congress and OCLC/Forest
fC d OCLC/F t
applications.
Press
• Search engine result
5. Library of Congress Subject Heading: :
improved.
US Library of Congress
6. WordNet: Pronceton University.
George Miller
7. CYC Ontology :CYC Corporation
LCSH : A quick look into its history
• The Library of Congress was established in 1800 and rebuilt
after the War of 1812 using Thomas Jefferson's personal
library, which he had sold to Congress. Along with the seven
thousand volumes came the classification method that
thousand volumes, came the classification method that
Jefferson had personally designed, using forty four classes
(Wynar 1985,403).
• Differing from Dewey, Dr. Putnam and his Chief Cataloguer,
Charles Martel, chose to use subject specialists to compile
each different schedule
each different schedule
16
2/22/2011
Thesaurus structure
• Concept‐term relationships
• Conceptual structure
• Semantic analysis and facets
• Hierarchy
How to convert thesaurus to ontology?
Models and modelmaking (May Subd
Geog)
[TT154]
UF Model‐making
Hierarchical Relation Models and
Modelmaking modelmaking
BT Handicraft
Manual training
Miniature objects
RT Modelmaking industry Express Partitive
SA subdivision Models under types of
SA subdivision Models under types of Is‐A
Is A Is A
Is‐A
Objects. e.g. Automobiles‐Models; R l ti
Relations
Machinery‐Models; and phrase
Headings for types of models, e.g.
Wind tunnel models
NT Architectural models
Engineering models Architectural
Geological modeling Handicraft
Geometrical models models
Historical models
Hydraulic models
Hydrologic models
Mannequins (Figures)
Miniature craft
Modelmakers
Query refinement
Models (Patents)
Models (Patents)
Paradigms (Social sciences)
tools for the searcher
tools for the searcher
Pattern‐making
Relief models
“is‐a” relations. Terms are
Ship models
Simulation methods
represented as classes
Surfaces, Models of
Wind tunnel models
BT = Superclass
Zoological models NT = Inverse relationship
Allow hierarchical (all classes having
navigation narrower meaning)
Feature: BT & NT Relationships
17
2/22/2011
How to convert thesaurus to ontology?
Associative Relation
Models and modelmaking
(May Subd Geog)
[[TT154]] Models and
UF Model‐making RT relations express
RT l ti modelmaking
Modelmaking any of a range of
BT Handicraft relations
Manual training hasRelatedTo
Miniature objects
RT Modelmaking industry
Modelmaking
RT & ‘See Also’ =
RT & ‘S Al ’ i d
industry
related terms (not
synonym or BT or NT) •Associative Relation
RT = any relation that does
not fall under BT/ NT
•Terms class related to
another class
How to convert thesaurus to ontology?
Context
Programming (Mathematics)
[QA 402.5]
UF Goal programming
p g g Mathematics
Mathematical
Programming
inTheContextOf
BT Algorithm
Functional equations
Mathematical
optimization Programming
Operations research
Context
•Term Association –
Increases the chances for represented as the
Headings serendipitous discovery of property ‘in the context
•Accompanied by a interesting terms of’
term in (…..)
18
2/22/2011
How to convert thesaurus to ontology?
Models and modelmaking Preferential Relation
(May Subd Geog)
[TT154] Model‐ making
UF Model‐making
UF Model making USE/UF relationship
USE/UF relationship
Modelmaking hasSynonynm
BT Handicraft
Manual training
Miniature objects Models and
modelmaking
Implies 2 terms are hasSynonynm
equivalent USE/UF
Modelmaking
Facilitate non expert uses to locate Represented as the
concepts they are searching ‘individual’ or ‘instance’
•Group all equivalent terms.
From Thesaurus to Ontology: Projects
Antique
Antique
Furniture
Ontology
Agricultural
Chinese
gy
Ontology
Travel
Travel
Service/
Domain
Concept
Ontology
Server
19
2/22/2011
Overview Project
Antique Furniture Agricultural Ontology Chinese Travel
Ontology Service/ Concept Server Domain Ontology
Antique Furniture Ontology
20
2/22/2011
Antique Furniture Ontology
1. Treated the main terms as concept names in
the knowledge base
the knowledge base.
– The full AAT hierarchy was converted into a
hierarchy of concepts, where each concept has a
label slot corresponding with the main term in
AAT and a synonyms slot where alternate terms
are represented.
– The knowledge base is represented in RDFS by
constructed an RDFS browser to inspect and
browse the hierarchy (Figure 2)
Antique Furniture Ontology
Figure 2
21
2/22/2011
Antique Furniture Ontology
2. Augment a number of concepts with additional
slots and fillers.
– For example, concepts representing a style or
period were augmented with slots time period
from, time period to, general style and region.
The values for these slots were partly derived
using explicit tables of periods, and partly by
using the intermediate concepts in AAT.
3. Add knowledge about the relation between
3 dd o edge abou e e a o be ee
possible values of fields and nodes in the
knowledge base (WordNet, Special purpose
documents)
Agricultural Ontology Service/ Concept
Server
22
2/22/2011
Agricultural Ontology Service/ Concept
Server
BT & NT relationships
Food Safety
dS f
isAbout
Quality
'is‐a' relationship by
default and can be re‐
interpreted to relate to
others with as much as
possible when needed.
BT & NT relationships
Agricultural Ontology Service/ Concept
Server
RT relationship
Contamination
canCaused
Food Safety
can be interpreted back to the
more specific relationships
RT relationship
23
2/22/2011
Agricultural Ontology Service/
Concept Server
• Use lexicalization to represent structure in multiple language
• Each term (lexicalization word) that describe concept in a
Each term (lexicalization word) that describe concept in a
specific language is modeled as instance of a concept
Chinese Travel Domain Ontology
24
2/22/2011
Chinese Travel Domain Ontology
– According to Xing et al. (2009), there are four key
elements in the development of the ontology:
• Terms
• Hierarchies
• Semantic Network
• Ability reasoning
– However, participation in the development of an
ontology is a thesaurus on two elements of the term and
hierarchy.
hierarchy
Chinese Travel Domain Ontology
• Term = concepts of the domain formally (foundation of
building ontology)
• Use thesaurus to standardize terms.
But…
• Thesaurus cannot completely describe some specific
areas in details.
• According to domain knowledge, we must add new terms
According to domain knowledge we must add new terms
to describe some important concepts that don’t exist in
thesaurus
25
2/22/2011
Chinese Travel Domain Ontology
• So, they define 2 types of terms
Conceptual term = Conceptual
Abstract term = Abstract class
class
(From domain expert)
(From thesaurus)
Chinese Travel Domain Ontology
• On the base of terms, hierarchy is another important element
of ontology.
Among different object class refer to
A diff bj l f
inheriting relationship
(is‐a, kind‐of, part‐of)
• Hierarchy
Among different classes refer to combination
relationship
(intersection, union, inverse set,
complementary set of other classes)
26
2/22/2011
Chinese Travel Domain Ontology
• By this way, term (concepts) can connect
t th b hi
together by hierarchy.
h
• Thesaurus as vocabulary table also has
hierarchy.
• After a little change and process, it can be
used in ontology
used in ontology.
Chinese Travel Domain Ontology
Change of Thesaurus Hierarchy
To change the hierarchy
g y
T di id
To divide new hierarchy according
hi h di
according to domain knowledge.
to domain knowledge in which the
thesaurus did not divide specifically
27
2/22/2011
Chinese Travel Domain Ontology
• Lastly, a part of hierarchy relationship of the ontology in China
travel field.
Overview Method
Antique Furniture Agricultural Ontology Chinese Travel Domain Ontology
Ontology Service/ Concept Server
28
2/22/2011
Why choose thesaurus in building ontology?
– To standardize terms (Xing et al., 2009)
• The standardized terminologies and professional division of
thesaurus can satisfy the requests of clarity, completeness, and
coherence on ontology.
• Terms will have extendibility so that you can continue to add new
term without changing original terms.
– To save times in building ontology because it consume a
lot of times if we fully depend on domain expert.
But…
Why choose thesaurus in building ontology?
– we still need the domain expert to add more
attributes and relationships to the ontology
attributes and relationships to the ontology
because thesaurus ‐ lack of relationship (Lauser et
al., 2006 and Xing et al., 2009)
• Thus, we need thesaurus and domain expert in
building an efficient ontology
29
2/22/2011
Institutions/ Organizations involve in
Ontology Research
Creating Domain Ontology
1. Select 1 thesaurus with a list of domain
coverage.
2. Select the related part Æ transfer to the
ontology.
3. More attribute/relationships can be added
based on consultation with a domain
based on consultation with a domain
specialist
30
2/22/2011
Chinese Agricultural Thesaurus
• Data
transferred to
Access
Database
• Java
J
programmer
Fig 2 used to
generate a Fig 3
KAON file from
the database
Chinese Agricultural Thesaurus
Fig 4
The resulting RDFS file has all the relations of a thesaurus
31
2/22/2011
I‐ES (Islamic Extraction System)
1 SSelect LCSH together
1. l t LCSH t th
with their related part
2. Use LCC‐BP to select
related parts and add
more attributes.
3. Use other reference
sources‐ Index
Islamicus
4. WIKI
5. Domain expert
I‐ES™
Select LCSH together with their related part
LCSH…
• Refer LCSHÆPhrase:’hijrah’
Class number
Related
Term
32
2/22/2011
I‐ES
I‐ES™
LCSH…
• Output after referring to LCSH
I‐ES
I‐ES™
Use LCC‐BP to select related parts and add more
attributes
attributes.
LCC…
33
2/22/2011
I‐ES
I‐ES™
LCC…
• Output after refer to LCC (BP 77.5)
I‐ES
I‐ES™
Use other reference sources‐ Index Islamicus
34
2/22/2011
I‐ES
I‐ES™
Index Islamicus…
• Output after refer to Index Islamicus
I‐ES
I‐ES™
WIKI
35
2/22/2011
I‐ES
I‐ES™
WIKI…
• Output after refer to WIKI
I‐ES
I‐ES™
Domain expert
Name: Prof Dr. Jawiah Dakir
Position: 1) Deputy Director at Institut Islam Hadhari, Universiti Kebangsaan
Malaysia
2) Lecturer at Jabatan Usuluddin dan Falsafah, Fakulti Pengajian Islam,
Universiti Kebangsaan Malaysia
36
2/22/2011
IES Search Results
IES Search Results
37
2/22/2011
Uniqueness of
I‐ES:
Malay Web Document
English Web Document
38
2/22/2011
Arabic Web Document
HikMas™
HikMas™
Holistic Knowledge Management System
39
2/22/2011
IGR ‐ Ontology keyword searching
result
Conclusion
• The goal of the Semantic Web initiative is to
annotate large amounts of information
annotate large amounts of information
resources with knowledge‐rich metadata.
• Building ontologies for large domain such as
agriculture, medicine, occupation, education
or arts is a costly affair.
• However, many domain thesauri have been
H d i th ih b
built can be a basis for the construction of an
ontology.
40
2/22/2011
Conclusion
• A thesaurus should satisfy in number of criteria:
¾It should have a strict subclass, superclass
¾It should have a strict subclass superclass hierarchical
hierarchical
structure
¾It should be base on unique concepts rather than on
natural‐language terms.
¾It should be representable in a format that is
compliance with emerging web standards.
¾in ontology construction, additional knowledge
should be added to the basic hierarchical structure of
h ld b dd d h b i hi hi l f
concepts derived from the thesaurus.
¾This knowledge can comes from different sources:
LCSH, LCC, AAT, WordNet, SIC, NAIC etc.
41