Dual-Stage Social Friend Recommendation System Based on User Interests

Abstract

The use of online social network (OSN) has become essential to humans' lives whether for entertainment, business or shopping. This increasing use of OSN motivates designing and implementing special systems that use OSN users' data to provide better user experience using machine learning and data mining algorithms and techniques. One system that is used extensively for this purpose is friend recommendation system (FRS) in which it recommends users to other users in professional or entertaining online social networks. For this purpose, this study proposes a novel friend recommendation system, namely Hybrid Friend Recommendation (FR) model. The Hybrid model applies dual-stage methodology on unlabeled data of 1241 users collected from OSN users via our online survey platform featuring user interests and activities based upon which users with similar social behavioral patterns are recommended to each other. The model employs a variety of techniques including user-based collaborative filtering (UBCF) approach and graph-based approach friend-of-friend recommendation (FOF). The model offers unique solutions to common problems of FRS such as data sparsity by using dimensionality technique called non-negative matrix factorization (NMF) to create a dense representation of the collected data and reduce its sparsity as well as providing seamless integration with other FRSs. The evaluation of the hybrid FR model shows positive correlation of Pearson correlation coefficient (PCC) compared to the baseline used.