Credibility-based Social Network Recommendation: Follow the Leader

In Web-based social networks (WBSN), social trust relationships between users indicate the similarity of their needs and opinions. Trust can be used to make recommendations on the web because trust information enables the clustering of users based on their credibility which is an aggregation of expertise and trustworthiness. In this paper, we propose a new approach to making recommendations based on leaders’ credibility in the “Follow the Leader” model as Top-N recommenders by incorporating social network information into user-based collaborative filtering. To demonstrate the feasibility and effectiveness of “Follow the Leader” as a new approach to making recommendations, first we develop a new analytical tool, Social Network Analysis Studio (SNAS), that captures real data and used it to verify the proposed model using the Epinions dataset. The empirical results demonstrate that our approach is a significantly innovative approach to making effective collaborative filtering based recommendations especially for cold start users.