Social network enhanced collective recommendation
Wang Zhou, Jianping Li, Qingtao Xue · 2017
Aimed to address the data sparsity and cold start problem, this paper tries to improve the recommendation accuracy via exploiting the hidden social information from vast amounts of data, and proposes a general social network enhanced collective recommendation model using matrix factorization (SNERM), which measures the social relationship through user's historical behavior and the interplay between users, and conducts recommendation in each inferred social community respectively. The experimental results on large real world datasets demonstrate that our proposed method outperforms other state-of-the-art approaches, especially in recommendation accuracy and solving the cold start problem.