Collaborative Filtering Recommender System Based on Social Network

Abstract

In recent years, the use of social network services is constantly increasing. A social network service (SNS) is an individual-centered online service that provides means for users to share information and interact over the Internet. In a SNS, recommender systems supporting filtering of substantial quantities of data are essential. Collaborative filtering (CF) used in recommender systems produces predictions about the interests of a user by collecting preferences or taste information from many users. The disadvantage with the CF approach is that it produces recommendations relying on the opinions of a larger community (i.e., recommendations are determined based on what a much larger community thinks of an item). To address this problem, this article exploits social relations between people in a social network. That is, the recommender system proposed in this article takes into account social relations between users in performing collaborative filtering. The performance of the proposed recommender system was evaluated using the mean absolute error.