PERSONALIZED RANKING OF SEARCH RESULTS BASED ON SEMANTIC SIMILARITY BETWEEN QUERIES

With the rapid growth of the information available on the web as vvell as the ambiguity of typical user queries, it becomes difficult for web search engines to satisfy the user infhrmation need expressed by few keywords . The involved retrieval approaches provide the same results for the same keyword queries even though these queries are submitted by different users with different intentions. Personalized information retrieval is a promising way to resolve this problem by modeling the user profile by his general interests on the basis of search history and then integrating it in a personalized document ranking. In this paper we propose an approach for personalized ranking o.fsearch results. We examined the click history of each user and build a profile for every individual to identify his search interests. Jn particular, we used the topical ontology structure of URLs from ODP¹ (Open Directory Project) to determine the topics of interest for each query submitted by the user, we then measure the semantic similarity between user s current and past queries from the user profile to identify the topics that are relevant to his current query. Finally we utilize the similarity between topics to re-rank search results. Our experimental results show that the proposed approach can ejJectively capture user-specific query-dependent personalization and improve the accuracy of web search.