A Multilevel Semantic Document Classifier Based On SVM Integrated With Domain Ontologies

A multilevel semantic document classification system based on Support Vector Machine (SVM) in association with domain ontologies has been developed. The documents related to the scientific domains such as computer science and chemistry are treated as the test source. The classification results are more precise and fine grained when compared to the conventional methodologies. The sharpness of the classification has been found to be enhanced when the domain knowledge in terms of ontologies is integrated with SVM procedures. So the developed system provides the advantages of high generalization performance, prevention of over fitting, less computational complexity, high accuracy, and robustness. The use of automated identification of the semantic components derived from the domain ontologies enables the system to provide semantically rich classification results. Keywords Document Classification, Support Vector Machine, Ontology, Semantic components, Text Mining.

• Type: Journal
• Year: 2008
• Where: JCIT
• Authors: Vijayasundaram Uma, Punnaivanam Sankar, Gnanasekaran Aghila
• Keywords: Document Classification, Domain Ontologies, JCIT 2008, Support Vector Machine