A novel recommendation system approach utilizing social network profiles
Timo Kahara, Keijo Haataja, Pekka Toivanen · 2013
In this paper, a comparative analysis of recommender systems (RSs) to mitigate Cold-Start Problem (CSP) is provided. In addition, a novel way to utilize social network profiles for recommendations is proposed. The proposed DF (Demographic Filtering) and COBF (Community-Based Filtering) based hybrid approach is designed to facilitate user registration process in sites employing RSs and to alleviate CSP, thus enabling the design and implementation of more efficient and accurate RSs. In our proposal, the traditional way of user profile creation is eliminated and it is replaced by the utilization of social network profiles. The purpose of this paper is to help RSs' designers to better understand CSP and to provide some solutions for it as well as to simplify the design and implementation process of Social Recommendation Systems (SRSs), also referred to as COBF-based RSs. Moreover, some new ideas that will be used in our future research work are proposed.