WikiOnto: A System for Semi-Automatic Extraction and Modeling of Ontologies Using Wikipedia XML Corpus

Abstract

This thesis investigates the problem of extracting and modeling ontologies to be used in a variety of scenarios using the Wikipedia XML Corpus as its source. The present research in terms of creating, extracting and learning ontologies con ne mostly to the domain of plain text and this has left a void in the area of semi-structured data, even when a large number of communications in the current world are being carried out using these methods of data interchange. We model an ontology extracting and modeling environment where the user is able to quickly model an ontology from the Wikipedia XML Corpus and is then able to create ontologies from the concepts that are available in those documents. A three-tiered architecture is presented where the user takes advantage of the structure of the documents, a word clustering algorithm and Natural Language Processing techniques to re ne the ontology he/she is creating. Finally, we enable the graphical representation of the ontology being developed, giving the user unlimited freedom to make any needed changes to the ontology that is taking shape.