Unsupervised approach for shallow domain ontology construction from corpus

Subhabrata Mukherjee, Jitendra Ajmera, Sachindra Joshi

Abstract: In this work we propose an unsupervised approach to construct a domain-specific ontology from corpus. It is essential for Information Retrieval systems to identify important domain concepts and relationships between them. We identify important domain terms of which multi-words form an important component. Our approach identifies 40% of the domain terms, compared to 22% identified by WordNet on manually annotated smartphone data. We propose an approach to construct a shallow ontology from discovered domain terms by identifying four domain relations namely, Synonyms ('similar-to'), Type-Of ('is-a'), Action-On ('methods') and Feature-Of ('attributes'), where we achieve an F-Score of 49.14%, 65.5%, 65% and 80% respectively