Authors: Xiaoyi Deng; Feifei Huangfu
Addresses: Business School, Huaqiao University, Quanzhou 362021, China ' College of Foreign Languages, Huaqiao University, Quanzhou 362021, China
Abstract: Social networks make users more dependent on online information regarding purchasing decision making. Networks which make users more dependent on online information regarding purchasing decision making. Therefore, social network information can be utilised to improve the performance of recommender systems that aim to mitigate information overload and provide users with the most attractive and relevant items. To improve recommender systems by incorporating social network information, this paper exploits multi-sourced information to predict ratings and make recommendations. An improved collaborative topic regression model that incorporates social trust, in which users' decisions regarding ratings are affected by their preferences and the favours of trusted friends, is proposed. In addition, an approach to calculating the maximum a posteriori estimates is proposed to learn model parameters. Empirical experiments using two real-world datasets are conducted to evaluate the performance of our model. The results indicate that the proposed model has better accuracy and robustness than other methods for making recommendations.
Keywords: collaborative topic regression; matrix factorisation; social trust; trust propagation; recommender system.
International Journal of Information Technology and Management, 2019 Vol.18 No.2/3, pp.182 - 194
Received: 26 Jun 2017
Accepted: 30 Jun 2018
Published online: 10 May 2019 *