Social recommendation algorithm based on stochastic gradient matrix decomposition in social network

• Published in: Journal of ambient intelligence and humanized computing Vol. 11; no. 2; pp. 601 - 608
• Main Authors: Zhang, Tian-wu, Li, Wei-ping, Wang, Lu, Yang, Jie
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.02.2020; Springer Nature B.V
• Subjects: Accuracy; Algorithms; Artificial Intelligence; Collaboration; Computational Intelligence; Decomposition; Empirical analysis; Engineering; Factorization; Methods; Original Research; Preferences; Recommender systems; Regularization; Robotics and Automation; Social networks; User Interfaces and Human Computer Interaction; Web 2.0; Social network; Matrix decomposition; Stochastic gradient; Recommendation system
• ISSN: 1868-5137; 1868-5145
• DOI: 10.1007/s12652-018-1167-7

The revenue of an e-commerce system is affected directly by the prediction accuracy of recommendation system. Although recommendation systems have been comprehensively analyzed in the past decade, the study of social-based recommendation systems just started. In this paper, aiming at providing a general method for improving recommendation systems by incorporating social network information, we propose a social recommendation algorithm based on stochastic gradient matrix decomposition in social network so as to improve the prediction accuracy. This paper considered the social network as auxiliary information, and proposed a matrix factorization based on social recommendation algorithm, which systematically illustrate how to design a matrix factorization objective function with social regularization. It constructed a matrix with the social network and the user scoring matrix, and proposed a stochastic gradient descent algorithm for matrix factorization. The empirical analysis on two large datasets demonstrates our proposed algorithm has lower prediction error, and is obviously better than other state-of-the-art methods.