Query Subtopic Mining from Search Log Data

Abstract

The World Wide Web has become an integral part of our lives. This large store of data is accessed through search engines. Hence, search engines play a very important role in accessing information. But what the search engines lack is an efficient mechanism to return data which is relevant to the users. The queries fired by users tend to be ambiguous and multifaceted. Deriving the most likely facet from the queries and thus presenting relevant results to the user is a difficult task. Consider a user fires a query which is a homonym, the user has to sift through this huge list of results to find the links relevant to his search. User’s past behavior is an effective parameter to determine the user’s probable search intent. The user is likely to be interested in a particular facet of a multifaceted query, if he has searched for it in the past. Our method deals with returning customized results and suggestions, to users based on their previous search history, in response to a query. For this we refer to the user’s search logs to obtain knowledge about his interests and based on this knowledge return customized and user relevant results.