Improving collaborative filtering recommender system results and performance using satisfaction degree and emotions of users

Article type: Research Article

Authors: Alhijawi, Bushra^{; *}

Affiliations: Computer Science, Princess Sumaya University for Technology, Jordan. E-mail: [email protected]

Correspondence: [*] Corresponding author. E-mail: [email protected].

Abstract: Recommender Systems (RSs) are provided as web 2.0 services that give support to electronic commerce. Today, social information has been used for improving the performances of RSs. One of the popular social networks is Facebook which recently developed a new reaction button. This button provides new opportunists to analyze and understand the user’s emotions and behavior. This paper proposes a new social-based RS that benefits from social information to improve the performance of the collaborative filtering. This RS depends on satisfaction degree and emotions of users. Each user can rate an item to express his satisfaction degree with this item. Also, the user able to express his feelings toward this item through the Facebook reaction button. The proposed algorithm is experimentally compared to alternative techniques. The results obtained shows that the proposed algorithm outperforms these algorithms in recommendation quality by 40% and performance by 29%.

Keywords: Recommender system, collaborative filtering, Facebook, social network, social information

DOI: 10.3233/WEB-190415

Journal: Web Intelligence, vol. 17, no. 3, pp. 229-241, 2019