Abstract: In recent years, using different trust models in recommender systems based on collaborative filtering, provides a trustable environment for transactions and interactions between vendors and vendees. Several trust based models were created based on rates given by users to different products and items in the past and also based on relational and popularity trust. In this study, a new recommender system is proposed by combining personal and group viewpoints. Personal viewpoints are computed according to rates given by the users. The On the other hand, user whose favorites and priorities are similar for choosing a website items can make a group together. In our method, group trust is computed based on users weights in a group. Eventually, our recommender system recommends by using personal and group trusts and combining these two viewpoints and adding entropy. The Our experimental results shows that our model increases performance and accuracy of prediction significantly in compare with other collaborative filtering models.
Afsaneh Khosravani and Hassan Shojaee- Mend, 2016. A New Recommender System Based in Collaborative Filtering Based on Personal and Group Trust for Offering Significant Suggestions to the Users. Asian Journal of Information Technology, 15: 2296-2305.