Semantics based information retrieval using conceptual indexing of documents

This paper proposes a technique for information retrieval using conceptual indexing of documents. A novel word sense disambiguation approach is applied to the set of input documents and the senses of the words are accurately determined using the senses present in the WordNet along with the contextual information present in the document. Once the senses are determined, the documents are indexed conceptually. The group of closely related synsets has been defined as a concept. The query is also conceptually disambiguated. Once the documents and the query are brought to the same format, retrieval of documents is performed and the results show improved effectiveness over other retrieval systems.