An improvised filtering based intelligent recommendation technique for web personalization

Personalization is an attempt at addressing a service provider´s desire to push additional information to users visiting their domains while at the same time restricting the flow of irrelevant recommendations. Web Personalization is viewed as an application of web mining and machine learning techniques for improved user satisfaction. The two commonly used methods of web personalization are Content-based filtering approach and Collaborative filtering approach. However, the most successful recommender system for web personalization is the collaborative filter since the content based filter has its own drawbacks. Apart from these, the most complicated problem of conventional collaborative filtering is the shilling effect. Item based algorithms avoid this main backlog in the conventional collaborative filter by reducing the effect of user similarities. Thus, user´s neighbourhood interference is considerably reduced and the item based prediction is given more priority. In this paper, we propose an intelligent recommendation system that utilises (1) Boosted item based collaborative filtering for the efficient rating of predicted items and (2) Association rule mining technique for making a personalised recommender system for the target user. This improves the overall web recommendation precision.