PolyU IR
 

PolyU Institutional Repository >
Computing >
COMP Theses >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/4271

Title: Ontology learning in Chinese for information search and management
Authors: Lim, Hon-yeung, Edward
Subjects: Hong Kong Polytechnic University -- Dissertations
Expert systems (Computer science)
Knowledge acquisition (Expert systems)
Ontology
Issue Date: 2011
Publisher: The Hong Kong Polytechnic University
Abstract: Ontology is an effective approach for representing knowledge in computer systems. It is an important technology for developing intelligent knowledge-based information systems. Many such ontologies representing different domains of knowledge have been developed in recent years. They are mostly created manually by ontology engineers and domain experts. This creation method is however inefficient and time consuming. Ontology learning is therefore a practical approach to support ontology engineers and domain experts in conceptualizing the knowledge of a particular domain. Techniques of ontology learning in recent research mostly concern on using texts as the learning source, as text data is a rich and direct source of human knowledge. This research proposes a comprehensive ontology based system framework called KnowledgeSeeker, which contains four different ontological components and processes that can be used to develop different kinds of ontology-based information systems. First, the framework defines an ontology representation model called Ontology Graph, which defines the ontology and the knowledge conceptualization model in a graphical format. Second, an ontology learning process that based on chi-square statistics is proposed for automatic learning an Ontology Graph from texts for different domains, as called Domain Ontology Graph (DOG). Third, it defines an ontology generation method that transforms the learning outcome to the Ontology Graph format for machine processing and also can be visualized for human validation. Fourth, it defines different ontological operations (such as similarity measurement and text classification) that can be carried out with the use of generated DOGs. This research focuses on Chinese text data and therefore we conduct experiments of the ontology learning process by using Chinese texts as the learning input. The experiment generated 10 DOGs as the Ontology Graph instances to represent 10 different domains of knowledge. The generated DOGs are then further used for an experiment of Ontology Graph based text classification providing performance evaluation. The experiment is able to achieve high text classification accuracy (with 92.3% in f-measure) over other text classification approaches by using the Ontology Graph based approach. The high performance in the experimental result reveals that the proposed Ontology Graph model, the ontology learning process, and the defined ontological operations are effectively developed. A commercial product that adopts the techniques of KnowledgeSeeker, called IATOPIA iCMS KnowledgeSeeker, with two real applications called 1) IATOPIA News Channel (IAToNews) and 2) IATOPIA Digital Asset Management System (DAMS) are presented to demonstrate the use of KnowledgeSeeker technique to develop intelligent ontology-based information systems.
Degree: M.Phil., Dept. of Computing, The Hong Kong Polytechnic University, 2011
Description: xv, 182 leaves : ill. ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577M COMP 2011 Lim
Rights: All rights reserved.
Type: Thesis
URI: http://hdl.handle.net/10397/4271
Appears in Collections:COMP Theses
PolyU Electronic Theses

Files in This Item:

File Description SizeFormat
b24250636_ir.pdfFor All Users (Non-printable) 21.46 MBAdobe PDFView/Open
b24250636_link.htmFor PolyU Users 162 BHTMLView/Open



Facebook Facebook del.icio.us del.icio.us LinkedIn LinkedIn


All items in the PolyU Institutional Repository are protected by copyright, with all rights reserved, unless otherwise indicated.
No item in the PolyU IR may be reproduced for commercial or resale purposes.

 

© Pao Yue-kong Library, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
Powered by DSpace (Version 1.5.2)  © MIT and HP
Feedback | Privacy Policy Statement | Copyright & Restrictions - Feedback