Trust dependence network for recommender systems

The recent invasion of social networks makes it inevitable to construct a trust network. Recommender systems at present uses advanced parallelism in web development. But attaining trustworthiness in such a system has been a challenging task in research work for several years. To overcome the issues in the existing work such as sparsity, scalability and to eliminate distrust users in the network, the proposed framework focused on some influential metric features such as similarity and centrality measures. Finally an optimized, trustworthy network is generated which is experimentally verified with the dataset from an online sharing community network Epinion.com.