iAGENT: A Novel and Intelligent Assistant to Personalised Search

This paper presents a novel approach to personalised search. By constantly learning and updating user profiles to the current needs of the user, relevant results are returned. The paper puts forward the concept of “iAGENT” which is an intelligent agent that assists a user to get relevant documents by modifying the query given by the user in accordance with the web pages previously visited. iAGENT has a novel method of maintaining two kinds of profile for the user. The first profile keeps track of the pages visited by the user and the second keeps track of the pages that were not visited by the user. The second profile is an added feature of iAGENT that keeps a backup of the websites that were irrelevant at some point of time, but may be required in the future. When a request to a page from the second profile is made, the overhead of following the entire procedure of querying the search engine is avoided. Besides, both the profiles are constantly updated to keep track of the user’s current interest. Experimental results show that the iAGENT is efficient enough to personalise the search results to an appreciable degree.