Building User Profiles Based on Preference Networks
J. Gnana Jayanthi, Bora Keerthana, K. S. Jayakumar · 2012
As the number of internet users grows rapidly, it is important to provide mo re relevant results for the individual user. It can be ach ieved by the agent that tracks and mon itors user's behavior. Users search behavior is stored in the form of profile.User's interest is analysed based on various factors like query,usage count,concept age etc.The proposed work in the profile construction includes following tasks,(i)Extraction of intents(set of keyword extracted fro m top k documents)(ii)Construction of preference network(iii)Creation and updating of the user profile.These profiles could be used to retrieve the prefered search results.Profile is organised in an hierarachial way so the particular concept along with the keywords can be viewed easily.User interest is tracked in a better way and it would imp rove the search results with more precision.