Finding similar patents through semantic expansion

Semantic information retrieval technique is widely applied in various research areas. However the key consideration is about selecting the most extensive external source which can be used for augmenting the query. In this research paper we have employed WordNet and Wiktionary as two external sources for expanding the query and finding the effect of it over the similarity models. We found that a combination of two external sources gives much improved result in comparison to single source expansion. We have further selected wu-palmer model to measure the efficiency of these combined models of expansion to find the effect on similarity in comparison to the tradition similarity cosine model.