Concept Based Query Recommendation

For a search engine, the challenge of finding relevant information from the web is becoming more and more difficult with rapid increase/change in content of the web. This difficulty further increases as queries submitted by users are general, imprecise, short and ambiguous. Relevance between user’s information need and documents returned by search engine is largely dependent on the query given by them. In this paper, we have proposed a method to facilitate users with query recommendations which are the concepts related to their information needs. In this work, we have extracted concepts from the web snippets and we have proposed two weight functions to measure the relevance between query and concepts. Related concepts with different meaning are selected and recommended as query suggestions. To evaluate our method, we have used a Google middleware for the extraction of concepts. We have estimated the relevance between the query and concepts using the proposed weight functions and compared with the support of the concepts as well as with the TFIDF approach using the standard information-retrieval metrics of precision and Mean Average Precision(MAP). We show that our approach leads to gains in average precision than the other existing approach for different type of queries. .