Intelligent user interest prediction using clock rate measures for efficient personalized web search

Abstract

The web search is the dominant activity of many web users, and newlinemost activities are performed with the support of the web. Quite a number of newlineweb techniques have been developed, but user expectations change all the newlinetime. A single user might have so many interests, but it is not necessary that newlinethe people s interest should be static. For example, a user might have interest newlinein reading topics related to and#8213;Data Miningand#8214;, but the next day his interest might newlinechange to read about and#8213;Multimediaand#8214;. For any web search engine, it is newlinenecessary to produce higher impact results to the user. So there is a need to newlineidentify the user interest and monitor how the interest changes. The user newlineinterest would be determined based on the frequency of visit and the time newlinespent on any topic. Due to changing user trends, it is not efficient since the newlineuser has higher expectations on the web results.The First approach is Intelligent Web Inference Model (IWIM).Initially, the constructed query URL will be connected and the query newlineprocessor will receive the HTTP result returned by the URL. The obtained newlineresult will be handed over to the rest of the components of the model thus newlineproducing most effective results. Based on the calculated state support value, newlinethe method calculates interestingness likelihood and finally makes references newlineto the prediction. The web inference model performs the change of interest newlineoccurred in multi-user search history. From the weblog available, for each newlineuser, interestingness of user at different time window and the change of newlineinterest both are identified. From the recognized interest, the inference model newlineidentifies the probable future interest of user and groups the similar users with newlinethe same interest hence collecting the pages visited by them to provide more newlinerelated results to the user. newline newline