Recommender System in Social Networks using Fuzzy Logic

Today, the importance of recommendation systems (RSs) has led to extensive activities and research efforts to improve the accuracy of the recommendations in different applications of social networks. Because various uncertainties exist in computing similarity between users and movies, it is an important issue to achieve high recommendation accuracy. This paper proposes a fuzzy logic-based recommender system (FRS) using Pearson correlation in social networks (SNs). The proposed scheme measures the similarity between users and movies by utilizing users’ interests based on labels, friendship, and group memberships for accurate recommendations in social networks. This paper defines a novel item prediction scheme that makes use of both user-to-user and item-to-item similarities. Experimental results show the effectiveness of the proposed scheme in improving prediction accuracy.