SIVA KALYANI,

Abstract — Collaborative filtering (CF) is a very important and common technology for recommender systems. Recom-mender systems are evidenced to be valuable means that for internet on-line users to deal with the data overload and became one amongst the foremost powerful and common tools in e-commerce. However, current CF ways suffer from such issues as knowledge sparseness, recommendation quality and big-error in predictions with lack of user privacy. There are 3 common approaches to determination the suggestion problem: ancient cooperative filtering, cluster models, and search-based ways and a completely unique rule to advocate things to users supported a hybrid technique. Initial we have a tendency to use cluster to create the user clusters supported the similarity of users. We’ve got taken users look history for similarity calculation. Second we have a tendency to be getting to realize the things that are powerfully related to one another by victimization association rule mining. Finally we'll be victimization these robust association rules in rec-ommendation of things. In ordered to supply the protection we have a tendency to used onion routing algorithms