Read My Research Experience on Ontology text version

SEMANTIC-BASED KNOWLEDGE MANAGEMENT TOOLS 3

My Research Experience on Ontology Optimizing rules and the Q-square Knowledge representation framework

Dr.Sasiporn Usanavasin: Advisor

Krich Intratip: PhD.Student

Outline

Introduction to my research trip Optimizing rules (Problems ­ Ideas ­ Method) Introduction to my current works

Ideas & Backgrounds Case Study

Research instrument & Ontology design Clear in concept and instance for well-form design Questions & Suggestions

My Research Trip

· Constructive Simulation (The best of Army research award'2006) · Grid Computing · Military Decision Making Process using Semantic Web · Ontology Improvement

­ Optimizing Rules (KICSS2010)

· Knowledge Representation using Ontology Based on Research Findings

­ A Development Framework for Qualitative and Quantitative Knowledge Based on Research Findings using Ontological Representation

http://itpe.siit.tu.ac.th/kicss2010/front/show/accepted-papers

Step-wise Approach for Improving Ontology using Optimizing Rules

Krich Intratip and Sasiporn Usanavasin Faculty of Information Technology, Sripatum University, Bangkok, Thailand

Semantic Web was introduced

"The Semantic Web is an extension of the current web in which information is given Well-defined meaning, better enabling computers and people to work in cooperation."

Berners-Lee at the Home Office, London, 2010

[1] T. Berners Lee, J. Hendler and O. Lassila. The Semantic Web. Scientific American, May 2001.

by Krich Intratip

Semantic Web was introduced

"The semantic web is designed to smoothly interconnect personal information management, enterprise application integration, and the global sharing of commercial, scientific and culture data. We are talking about data here, not human documents." Berners-Lee at the Home Office, London, 2010

The semantic web : an interview with Tim Berners Lee, Consortium Standard Bulletin, 2005. http://www.consortiuminfo.org/bulletins/semanticweb.php

by Krich Intratip

Problems

· Most of them are poor design at least one problem

­ Domain knowledge is in both the ontology and its programming code, hard to maintenance ­ One fact is in many places, hard to maintenance ­ Ontology is also bigger than usage, lack of performance ­ Other nodes or instances are also in the ontology, although it has the best one for the result, lack of performance

· However, the Semantic web application can go on running (by program solving)

by Krich Intratip

Example of composite value

Black t-shirt, color of t-shirt is black

Black color

Label

Black box, color of box is black

Red Box, color of box is red

by Krich Intratip

Black Label Red Label Product name

Example of composite value

Black t-shirt, color of t-shirt is black

Black color

Label

Composite value

Black box, color of box is black

Non-composite value

Red Box, color of box is red

by Krich Intratip

Black Label Red Label Product name

Example of composite value

Black t-shirt, color of t-shirt is black

Black color

Label

It is a design problem issue, not only programming problem issue.

Black box, color of box is black

Red Box, color of box is red

by Krich Intratip

Black Label Red Label Product name

Some domain knowledge are thrown in Programming area

Composite value "Black box"

"Black box"

What does it mean?

Separate to

Programming area

for

What is the box color?

Black

for

Box

Ontology area

What is the item?

by Krich Intratip

*

Got Ideas

Ontology design

Ontology improvement

Quality of ontology

by Krich Intratip

Our proposed method

· Step-wise approach for improving for ontology design ­ 4 optimizing rules

· Remove composite-values to optimize the maintenance · Remove one fact in many places to optimize the maintenance · Remove unused class to optimize the performance · Remove unnecessary class to optimize the maintenance and performance

"Optimize both the maintenance and performance"

by Krich Intratip

Scenario

· student ID, student name, admission year, major, level, and gender. · If one of the student record has values as `50560073', `Mr.Krich Intratip', `2550', `IT', `Ph.D.', `male'

by Krich Intratip

Scenario

:shows composite-values problem

· student ID, student name, admission year, major, level, and gender. · If one of the student records has values as `50560073', `Mr.Krich Intratip', `2550', `IT', `Ph.D.', `male' · student name (e.g., Mr.Krich Intratip) is a composite value · student name should be separate to title, first name and last name

by Krich Intratip

Scenario

: shows one fact in many places

owl:Thing

Students studentID `50560073' studentName `Mr.Krich Intratip' admissionYear `2550' major `IT' level `PhD.' gender `male'

ScholarshipRecipients

studentName `Mr.Krich Intratip' scholarshipTpye 'RTA. Officer' amount '15%off' duration `3 years'

by Krich Intratip

Solution (a)

owl:Thing

Students studentID `50560073' studentName `Mr.Krich Intratip' admissionYear `2550'

ScholarshipRecipients

studentName `http://www.spu.ac.th/Ontology.owl#Mr.Krich_Intratip' scholarshipTpye 'RTA. Officer' amount '15%off' duration `3 years'

major `IT'

level `PhD.' gender `male'

by Krich Intratip

Solution (b)

owl:Thing Students studentID `50560073' studentTitle `Mr.' studentFirstName `Krich' studentLastName `Intratip' admissionYear `2550' major `IT' level `PhD.' gender `male' ScholarshipRecipients scholarshipTpye 'RTA. Officer' amount '15%off' duration `3 years'

by Krich Intratip

Solution

: shows unnecessary class problem

owl:Thing

owl:Thing Students studentID `50560073' studentTitle `Mr.' studentFirstName `Krich' studentLastName `Intratip' admissionYear `2550' major `IT' level `PhD.' gender `male' ScholarshipRecipients scholarshipTpye 'RTA. Officer' amount '15%off' duration `3 years' Students studentID `50560073'

studentTitle `Mr.'

studentFirstName `Krich' studentLastName `Intratip' admissionYear `2550' major `IT' level `PhD.' ScholarshipRecipients scholarshipTpye 'RTA. Officer' amount '15%off' duration `3 years'

by Krich Intratip

Conclusions

· Eliminating the design issues such as

­ composite values of instances, redundancy of attributes and unused class nodes

· Enhance the ontology design such that

­ it can better serve to the usage or business objectives of the system

· Reduce programming overhead · Make the maintenance of ontology easier

by Krich Intratip

Current works

· Q-square knowledge representation framework (on going) · Root cause problem solving

Q-square = Qualitative & Quantitative Research Findings

INTRODUCTION TO KNOWLEDGE AND ONTOLOGY

(KM Processes)

Structuring Knowledge

. , Ontology for Information System Design and Development, 28 2553

Ontology

Ontology ? Domain experts( )/Stakeholders? Intend to use

· Ontology · Design of the ontology

Ontology improvement

by Krich Intratip

Ontology

Natalya F. Noy and Deborah L. McGuinness , Ontology Development 101: A Guide to Creating Your First Ontology, Stanford University, Stanford, CA, 94305

http://protege.stanford.edu/publications/ontology_development/ontology 101-noy-mcguinness.html

by Krich Intratip

Problems

· Gap between knowledge extraction and ontology design impact to...

­ No idea to sketch well-form design ontology ­ Need to do ontology improvement ­ Stuck in programming code ­ Hard to maintenance ­ Lack of reusable

by Krich Intratip

· (scientific method) -> research methodology

­ Qualitative research ­ Quantitative research

by Krich Intratip

·

­ Ontology Grounded theory

·

­ Path Ontology Structural Equation Modeling

by Krich Intratip

Research Areas

Ontology(knowledge representation) design approaches

Knowledge extraction using research methodology

Knowledge (pattern of information)

Research methodology (SEM, GT)

Ontology design & improvement

Transformation approaches by Krich Intratip

CASE STUDY

31

A B

" ?"

by Krich Intratip

32

Knowledge extraction & Ontology design

TOOL FOR SOLVING

SIMPLE STEPS TO CREATE ONTOLOGY

33

Research Instrument Ontology Development

· Domain · Intend to use · ­ Concept class Concept ­ Attribute class Class Instances · constrain Ontology Data type attribute ·

by Krich Intratip

Some of ontology (re)-engineering processes (Knowledge extraction)

· Define topic area

­ ?

· Define domain specific

­ ?

· Define intend to use (Domain expert)

­ ?

· Breakdown into sub-domains/concepts

­ ?

­ Review literature (Consider reuse)

· Define indicators in each concept · Define indicator measurement · Define scale of the measurement

by Krich Intratip

D1

Specification

<= ( %)

Power of Prediction

<= ( %) D2

D4 <= ( %)

Sensitivity

<= , ( %)

Consistency

D3

*

by Krich Intratip

No.

()

(, )

()

(4 )

comment

(D1+D2+ D3+D4)/4

/ > 50%

*

* Relevance theory: http://www.phon.ucl.ac.uk/home/PUB/WPL/02papers/wilson_sperber.pdf

*

by Krich Intratip

Instrument & Ontology design

Domain

Concept 2

Att. 1

Int

Concept 1

Att. 2

Att. 3 Att. 4 Att. 5

String

Intend to use

by Krich Intratip

*

Principle of defining class and its relation

· Class ()

­ class node 2

· Concept node node

­ : Instance

· Attribute node Instance

­ : Instance

­ concept node

· Is-a, Part-of

­ concept node Attribute node Attribute-of

by Krich Intratip

Principle of defining instance and its relation

· Instance (data item)

­ : Attribute node

· One fact in one place · Atomic value

· Relation

­ Attribute node Instance-of ­ Instance Attribute node Has_???

· Instance Node re-structure ontology

by Krich Intratip

Class/Instance and their relation

parent

concept concept concept

Is-a

concept

Part-of

concept

Att-of

attribute

Ins-of

instance

child

Physical/Concrete VS Logical

INFERENCE ? QUERY & RULE

43

Inference and Decision

·

­ (m) ­ (kg)

· (Query)

­ BMI = (m) * (kg)

·

­

http://en.wikipedia.org/wiki/First-order_logic http://www.chulabook.com/description.asp?barcode=978974 0326960 by Krich Intratip

4 3 2 1 1 2 3 < 16.0 16.0-16.9 17.0-18.4 18.5-19.9 20.0-24.9 25.0-29.9 30.0-39.9 > 40.0

by Krich Intratip

http://www.thailabonline.com/BMI.htm

Ontology

· (Generic VS Specific) · Ontology (Taxonomy) Hierarchy First order logic (Unique) Ontology · Reuse Ontology Ontology

by Krich Intratip

Ontology

· 5=X+2 · How to describe X?

­ ­ ­ ­ ­ ­ X=3 X=5-2 X=Y+1 X = f(Y) + 1; f(Y) = 2 X = f(Y) + f(Z); f(Y) = 2, f(Z) = 1 X=?

Part-of

5

Part-of

X

2

· 5 = f(X) + 2

by Krich Intratip

Knowledge & Ontology development life cycle

Describe domain and intend to use

Inference over ontology

Instrument development

meaning

Practice based ontology design Build up domain knowledge Gather information

by Krich Intratip

Knowledge & Ontology development life cycle

Describe domain and intend to use

Inference over ontology

Instrument development

meaning

Practice based ontology design Build up domain knowledge Gather information

by Krich Intratip

Dr.Sasiporn Usanavasin

FB: Zimmantic lab

https://www.facebook.com/groups/zimmaticlab/

FB: SEM Study Lab

https://www.facebook.com/groups/SEMStudyLab/

Krich Intratip

Slide @ http://prezi.com/d113kizjskll/nectec-semanticseminar-3-my-research-experience-on-ontology/

Information

My Research Experience on Ontology

50 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

711125


Notice: fwrite(): send of 199 bytes failed with errno=104 Connection reset by peer in /home/readbag.com/web/sphinxapi.php on line 531