Boosting social networks in Social Network-Based Recommender System

Abstract

E-commerce companies have integrated services in their websites in order to attract new users or guarantee the customers' fidelity. In order to accomplish these aims, recommender systems were developed as a tool to assist people in their purchases. Although, these systems have provided many advantages, they suffer from some drawbacks such the sparsity problem and the cold start problem. In order to smooth out both problems some solutions have been proposed. One of them is the integration of social networks in recommender systems creating a new paradigm of recommender systems called the Social Network-Based Recommender System (SNRS). In order to receive recommendations, these SNRSs require users to have, or provide, suitable social networks. However, social networks in e-commerce companies are usually embedded in their websites, and thus, users may not know enough acquaintances there to provide suitable social networks to the SNRS. In this contribution we address this problem and we present a model that, by means of interpersonal attraction theories, assists users in finding candidates who can belong to their social network. That way, not only does this model make easier the use of SNRSs, but it also encourages the use of the embedded social network, becoming an additional tool to improve the customers' fidelity.