Learning Concepts, Taxonomic and Nontaxonomic Relations from texts

Abstract

This paper discusses the knowledge extraction process in an ontology learning system called Hasti. It exploits an automatic, hybrid, symbolic approach to acquire conceptual knowledge and construct flexible and dynamic ontologies from scratch. This approach starts from a small kernel and learns concepts, taxonomic and non-taxonomic relations and axioms from natural language texts. The focus of this paper is on extraction of concepts and conceptual (taxonomic and non-taxonomic) relations using linguistic and template-driven methods. In this paper, the author will first present a brief overview on ontology learning systems and then describing the life cycle for the ontology learning and building process in Haiti, the knowledge extraction process will be discussed in more details. At last the author will present some experimental results of implementation and testing the proposed model