With the rich and growing wealth of information on the internet the process of finding a specific piece of information may often become frustrating and time-consuming for users. In the previous research to learn user interest using ontological profiles, the web pages visited and its dwell time have been considered. But considering these factors alone is not enough to extract user’s interest accurately. In this paper, a hybrid personalized model of search engine based on learning ontological user profiles implicitly is presented. The aim of this paper is to optimize the search results to get more relevant information rather than simple keyword matching by user’s recent browsing history. The user browsing behaviour is studied initially to get his area of interest and the search results for different users are based on his area of interest.  In addition, these web pages are stored in user profile under positive and negative documents. Thus, a hybrid reranking algorithm that is based on the combination of different significant information resources collected from the reference ontology, user profile and original search engine’s ranking has been proposed.  

Keywords: search engine, personalization, user profile, ontology, information retrieval.