Semi Automatic Ontology Development for Web

Abstract

All semantic web applications have their roots on ontology. However, development of a ontology is a very costly task in terms of time and money. Therefore, has become a bottleneck for the fast growth of semantic web. This thesis investigates the problem of extracting ontologies from large number of web documents. Even though present research in terms of creating, extracting ontologies mainly focus on unstructured web documents and fully structured web doc- uments, this research focus on semi structured web documents since most of the web content are in semi structured format. This thesis introduces web usage patterns as a novel source of semantics for extraction of conceptual relationships together with web content min- ing. Web site authors and web designers could conceptually view the web site/web pages in di erent ways. Therefore, by using web content and also usage patterns web author's thoughts as well as web user's intensions can be incorporated in the ontology development process. Furthermore, the method- ology presented in this thesis also minimizes the usage of web site speci c characteristics in the ontology learning process which increases its applica- bility. A prototype was developed as a proof of concept. The comprehensive qualitative and quantitative evaluation carried out has shown positive re- sults. This solution is intended to be usable for transformation of large web document corpus to semantic web and also it could be expected to be used to develop cross domain ontologies. Furthermore, many types of sites can pro t from reorganization as semantic web sites and these type of websites could be bene ted by this project.