USING THESAURI FOR TERM-BASED COMPONENT RETRIEVAL: A DISTRIBUTED AND ADAPTIVE MODEL

Abstract

When performing a query against a software repository, a thesaurus mechanism can help the user overcome vocabulary and language barriers. Because of the difficulty and high cost of building and populating a thesaurus, however, such a mechanism is provided by few repositories. In this paper we proppose a distributed and adaptive thesaurus model, which can reduce the cost of providing a thesaurus facility in a repository, make the thesaurus grow faster, and improve the performance of the query. The thesaurus architecture and the lookup algorithm are described formally, and an elementary analysis is given.