Matrix Factorization with Explicit Trust and Distrust Side Information for Improved Social Recommendation

Abstract

With the advent of online social networks, recommender systems have became crucial for the success of many online applications/services due to their significance role in tailoring these applications to user-specific needs or preferences. Despite their increasing popularity, in general, recommender systems suffer from data sparsity and cold-start problems. To alleviate these issues, in recent years, there has been an upsurge of interest in exploiting social information such as trust relations among users along with the rating data to improve the performance of recommender systems. The main motivation for exploiting trust information in the recommendation process stems from the observation that the ideas we are exposed to and the choices we make are significantly influenced by our social context. However, in large user communities, in addition to trust relations, distrust relations also exist between users. For instance, in Epinions, the concepts of personal "web of trust" and personal "block list" allow users to categorize their friends based on the quality of reviews into trusted and distrusted friends, respectively. Hence, it will be interesting to incorporate this new source of information in recommendation as well. In contrast to the incorporation of trust information in recommendation which is thriving, the potential of explicitly incorporating distrust relations is almost unexplored. In this article, we propose a matrix factorization-based model for recommendation in social rating networks that properly incorporates both trust and distrust relationships aiming to improve the quality of recommendations and mitigate the data sparsity and cold-start users issues. Through experiments on the Epinions dataset, we show that our new algorithm outperforms its standard trust-enhanced or distrustenhanced counterparts with respect to accuracy, thereby demonstrating the positive effect that incorporation of explicit distrust information can have on recommender systems. ACM Reference Format:Rana Forsati, Mehrdad Mahdavi, Mehrnoush Shamsfard, and Mohamed Sarwat. 2014. Matrix factorization with explicit trust and distrust side information for improved social recommendation.

Trust in Social Media

“…Model-based recommender systems with distrust use model-based trust-aware recommender systems especially matrix factorization based models as their basic models [35,88,128]. Assume that U i is the k-dimensional preference latent factor of u i and V j is the k-dimensional characteristic latent factor of item j p j .…”

“…where d is a distance metric and`is a penalty function that assesses the violation of latent factors of users with trust and distrust relations [35]. Possible choices of`.z/ are the hinge loss functioǹ .z/ D max.0; 1 z/ and the logistic loss function`.z/ D log.1 C e z /.…”

“…of regularization terms introduced by RecSSN[128] is n jON j where jON j is the number of users with only distrust relations, and it is much smaller than the number of regularization terms in Forsati et al[35].…”

A configuration-based recommender system for supporting e-commerce decisions

“…products for an existing consumer and the value of existing products for a new consumer without using additional data, such as explicit ratings (Zigoris and Zhang, 2006;Kim et al, 2011), product taxonomies (Weng et al, 2008), customer reviews (Levi et al, 2012), or social media data (Forsati et al, 2014;Yu et al, 2014;Zhao et al, 2016). Since additional data are not available in all contexts and for all consumers and products, the cold-start problem is still a challenge for content-based and collaborative filtering techniques.…”

“…A change of preferences of those consumers that demand a recommendation from a collaborative filtering or a content-based system likely reduces recommendation accuracy 3 . These two issues prevent improvements in the recommendation quality of content-based and collaborative-filtering recommender systems even if these systems use efficient methods, such as matrix factorization (Forsati et al, 2014). Neither of these issues arises in MAVT-based recommender systems.…”