Position Based On Refining Aggregate Recommendation Assortment

Abstract

Recommendation systems are becoming necessary for individual user and also for providing recommendations at individual level in various types of businesses. Recommender system is a personalized information filtering technique used to identify desired number of items based on interest of user. The system uses data on past user ratings by applying various techniques. This techniques concentrate to improve accuracy in recommendations, with recommendation accuracy it is also necessary improve aggregate diversity of recommendation. In this paper, we proposed number of item ranking techniques and different ratings prediction algorithm to improve recommendation accuracy and aggregate diversity by using real-world rating dataset.