AUTOMATIC ONTOLOGY CONSTRUCTION APPROACH FROM WEB PAGES

Abstract

In the thesis, a new approach for automatic ontology construction from web pages is proposed. The proposed approach works as follows: first the triples of the document are extracted, next it utilizes natural language processing techniques to process the extracted triples, and finally it applies ontology design patterns before inserting triples into ontology. Our proposed approach is compared with the previous related work and the results show that we are closed to cover all the information in the documents correctly. Our system is better at reducing repeat rate where it do not insert the same object twice and do not repeat the relation between two concepts. Also our ontology is simpler representation than other work.