INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: [email protected]...

33
INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: [email protected] Documentation Research and Training Centre Indian Statistical Institute Bangalore- 560 059 Workshop on Ontology, 19 th – 21 st March 2008. DRTC, ISI, Bangalore

Transcript of INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: [email protected]...

Page 1: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

INTO:

A way to represent institutional knowledge base

by

Biswanath Duttae-mail: [email protected]

Documentation Research and Training CentreIndian Statistical Institute

Bangalore- 560 059

Workshop on Ontology, 19th – 21st March 2008. DRTC, ISI, Bangalore

Page 2: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Introduction

Today “I am feeling lucky”! Oh… my God!! “What is this?” Why cannot you give me the desired document?? Oh… I got it. Ontology is great.

Page 3: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

But,…Why Ontology??

To share common understanding of the structure of information among people or software agents (Musen 1992; Gruber 1993)

To enable reuse of domain knowledge

To make domain assumptions explicit

To analyze domain knowledge

Page 4: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Objective…To

To show how the knowledge of an informal domain can be organized and retrieve them meaningfully to serve our various purposes

To show how the traditional classification system can be put in use for organizing the Web information

Page 5: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Methodology

Derived a set of potential questions (what do you want?)

Analysed the questions

Categorised them

Identified and collected the concepts and built relations

Followed controlled vocabularies for standardization

Applied classification principles (faceted)

Modelled the Ontology using OWL-DL

Integrated with SW browser (faceted web browser)

Page 6: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Languages and Tools Used

OWL-DL

Protégé 3.4 beta – a free, open source ontology editor and knowledge-

base framework

Pellet reasoner: open-source java based OWL-DL Reasoner

(http://www.mindswap.org/2003/pellet/)

SPARQL (SPARQL Protocol And RDF Query Language)

Colon Classification (7th ed.) Scheme

Longwell faceted web browser

Page 7: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Few Queries

Who are the experts on ontology in ISI along with their publications, home page (if any) and their e-mail addresses?

What are the different courses offered by ISI centre wise?

Who are the faculty members of a (particular) unit?

Is Dr. Prasad of DRTC earlier worked for some other organization? If yes how many years he worked there and what was his job profile?

Currently who is the head of DRTC and his contact details?

etc,…

Page 8: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Methodology

Derived a set of potential questions (what do you want?)

Analysed the questions

Categorised them

Identified and collected the concepts and built relations

Followed controlled vocabularies for standardization

Applied classification principles

Modelled the Ontology using OWL-DL

Integrated with SW browser (faceted web browser)

Page 9: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Ontology Metrics

Figure: showing metrics of classes and properties

Page 10: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 11: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 12: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Teacher and Student Class

Page 13: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 14: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 15: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 16: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Instance Tree

1 2

Page 17: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 18: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.
Page 19: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Example

Query:

Who are the experts on ontology in ISI along with their publications, home page (if any) and their e-mail addresses?

Page 20: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Analysis of the Query

To identify the experts, we tried to search for those people who are working on/or has research interest on Ontology along with their publication related to ontology (optional). Publication might provide their expertise in that area. Some other factors might also be considered, such as what events/ conference/ workshop they attended in recent past, etc.

Page 21: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

SPARQL: SPARQL Protocol And RDF Query Language

SPARQL is a query language

It is a protocol for accessing RDF (Resource Description Framework) data

It is designed by the W3C RDF Data Access Working Group

SPARQL builds on previous RDF query languages such as rdfDB, RDQL, and SeRQL

Further information is available in: http://www.w3.org/TR/rdf-sparql-query/

Page 22: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Query using SPARQL

PREFIX a: <http://localhost/Institute#>

SELECT ?Topic ?Name ?Publication ?Homepage ?emailWHERE { ?person a:researchInterestIn ?ResearchInterest . ?ResearchInterest a:title ?Topic . FILTER regex(?Topic, "ontology", "i") . ?person a:name ?Name

OPTIONAL { ?person a:publication ?Publication ; a:homepage ?Homepage } ?person a:email ?email }ORDER BY ?Name

Page 23: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Result from Pellet Reasoner

Page 24: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

??

But…, is it possible for end users to form such type of query (SPARQL query, etc.) ?

If no, what would be the possible solution?

Page 25: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Longwell

Longwell is a web-based RDF-powered highly-configurable faceted browser

It is the part of SIMILE project of MIT

It mixes the flexibility of the RDF data model with the effectiveness of the faceted browsing UI paradigm

It enables to visualize and browse any arbitrarily complex RDF dataset

It allows to build a user-friendly web site out of our data within minutes and without requiring any code at all

More information available on: http://simile.mit.edu/wiki/Longwell

Page 26: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

“INTO” with Longwell

Easy to integrate OWL Ontology with Longwell Web Browser

Knowledge base is easy to update

Capable of loading files that end with .rdf, .rdfs, .owl, .n3 and rss (Longwell assumes RSS 1.0 which is RDF; other versions will generate errors)

Page 27: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Longwell Faceted Web Browser

Page 28: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Longwell Faceted Web Browser

Page 29: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Longwell Faceted Web Browser

Page 30: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Observation

It is always better to go for modular based ontology

Conceptually draw your model first and then go for building ontology

Building ontology is quite expensive

No two ontologies designed by different people would be the same

There is no standard practice for building ontology

It's a biggest challenge to provide user friendly semantic search interface to end users

Page 31: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Wrap-up

Problems

Way to solve the problems

Why Ontology?

Methodology

Searching

Integration of INTO with a faceted web browser

Observations

Application

Page 32: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Reference (1)

Gruber, T.R. (1993). A Translation Approach to Portable Ontology Specification. Knowledge Acquisition 5: 199-220

Ranganathan, S. R. (1967). Prolegomena to library classification. Ed 3. Asia Pub House, Bombay

Ranganathan, S. R. (1987). Colon classification. SRELS, Bangalore

Bhattacharyya, G. (1979). POPSI: Its fundamentals and procedure based on a general theory of subject indexing language. Library Science with Slant to Documentation. V 16 (1)

Musen, M.A. (1992). Dimensions of knowledge sharing and reuse. Computers and Biomedical Research 25: 435-467

Page 33: INTO: A way to represent institutional knowledge base by Biswanath Dutta e-mail: dutta2005@gmail.com Documentation Research and Training Centre Indian.

Noy, N. F. and McGuinness, D. L. Ontology Development 101: A Guide to Creating Your First Ontology. http://protege.stanford.edu/publications/ontology_development/ontology101-noy-mcguinness.html

Protege (2000). The Protege Project. http://protege.stanford.edu

Flamenco. http://flamenco.berkeley.edu/

Reference (2)