IMSC Ontology Manager :Webbased Ontology Management System

1
IMSC Ontology ManagerWeb-based Ontology
Management System

• CSCI586 Project Group 3
• Sang-Soo Sung
• Rana Goyal, Jinal Jhaveri, Imran Naqvi,
  Sheng-Chih Lo, Ankur Verma
• Department of Computer Science
• University of Southern California

2
Outline

• Project Overview
• Distinguishing Properties
• IMSCOM Flowchart
• IMSCOM Demo
• IMSC Ontology Manager Architecture
• IMSCOM-API
• Milestone
• Team Organization
• Risk Management
• Future Work

3
Project Overview

• Current Web-compatible languages and technologies
  
• fail to allow a solid semantic expression
• do not provide enough support for other Web
  technologies
• have too simplified models
• lack the level of semantic information
• IMSC Ontology Manager
• examine, modify, construct and integrate
  ontologies dynamically
• richer semantic representations

4
Distinguishing Properties

• A solid and efficient method of building
  ontologies
• A human-oriented authoring interface
• A simple and accessible environment
• Multi-user authoring
• A publishing and security mechanism

5
IMSCOM Flowchart
Login
Show authors ontology lists
Choose ontology
?success
Yes
Start
no
Construct ontology
Create authors ontology
Edit authors ontology
Delete authors ontology
Create
Save
Edit
Ontology Management System
Concept
Delete
Publish
Attribute
Create
Ontology Manager
Schema Generator
Edit
Instance
Delete
6
IMSCOM Demo
httpimsc-dmim.usc.edu7777
7
IMSCOntologyManager Architecture
8
IMSCOM-API

• IMSCOM-API
• Easy to construct ontologies as a hierarchical
  structure
• Convenient to maintain ontologies on the server
  as well as on the clients.
• Efficient to augment ontologies with less time
  and less memory.

IMSCOM-API Data Structure
9
IMSCOM-API (Cont.)
Hash Table
Product Concept Object
ParentID None Operation None Predicate
None Children Book, Computer Attribute price
KEY
VALUE
1
Product
3
Book
Book Concept Object
ParentID 1 Operation None Predicate
None Children None Attribute ISBN, Title,
Author
6
Computer
Computer Concept Object
ParentID 1 Operation None Predicate
None Children None Attribute Vendor
10
Milestone
11
Team Organization

• Project Manager
• Sang-Soo Sung
• Team Management
• System Design
• DB Connection Pool
• DB Schema Design
• JSP Server-side programming

• GUI Developers
• Ankur Verma
• Ontology graph design programming
• GUI code integration
• Jinal Jhaveri
• GUI API design programming
• Attribute part programming
• JSP programming
• Imran Naqvi
• Ontology hierarchy design programming
• Menu design
• Concept part programming

• API Developers
• Sheng-chih Lo
• IMSCOM Ontology Storage design programming
• IMSCOM-API design programming
• Rana Gaurav Goyal
• XML Schema Generator programming

12
Risk Management

• Large size of memory requirement
• Shrink the applet size as small as we can
• Large size of data transmission between the
  server and clients
• Logs? XML-RPC?
• Send the serialized hash table as a character
  string
• Hard to handle huge size of programming code
• Applet around 5,000 lines
• IMSCOM-API around 3,000 lines
• DB Connection pool around 4,000 lines
• JSP and HTML around 1,400 lines
• Limited time
• Average 3 hrs for the group meeting per week
• Average 9 hrs for the individual work per week

Total Around 13,400 lines
13
Future Work

• Need a powerful Semantic Web language
• Support Triple or Quad?
• Enhance IMSC Ontology Manager to learn new
  concepts, attribute, and patterns from meta data
• Facilitate ontology analysis and concept mining
• Provide ontology matching and integration
  mechanisms at the semantic level

14
Thank you

• Question? Comment?