Improving Web Page Access Prediction using Web Usage Mining and Web Content Mining

Many times web users face the problem of information overload because there is huge collection of resources on the web. So there is a need to identify the web users' behaviors on web sites using web mining for improving the user's experiences on these sites. Prediction of the web page that might be visited by the web user is very important because this knowledge can be used for pre-fetching of web pages, recommendation or personalization of the web page for that user or group of users. By only considering the URLs of visited pages, it is not possible to capture the intention of the user so content of the web pages is also needed to capture the intention of the user by using the semantics which is present in the web pages. For current user, the next web page that might be visited by the user is predicted by searching the most similar cluster and then searching the most similar sessions in that cluster. It's not possible to store or manage all previous sessions and prediction time is increased due to comparison with all the sessions if the cluster has many sessions. This work proposes the strategies for discarding the old sessions for storing new sessions. It chooses the sessions which are going to be kept for prediction so that the prediction time is reduced with improved accuracy.