Domain Sensitive Recommendation using both User Item Subgroup Analysis and Social Trust Network

Abstract

Collaborative Filtering is a conventional technique and a booming Concept in recommendation systems to handle with overload information in the real world. In which the preference of a user on an item is predicted on the strength of the preferences of other users with homogeneous interests. CF attains relationships in between the users and suggests the items to other users. This paper introduces Social matrix factorization, to establish the rating matrix for the user and item. In which everyone convey their ratings on some items beyond creating social connection with other users. Domain clusteringis used for find the domains. The domain clustering model is rendered to make full use of the high- rated users and items are gathering them into groups.