Trust Based Recommender System

摘要

In the community we rely on recommendations from our trustworthy acquaintances to recommend products based on their experience. Recommender systems use opinions of members of a community to help individuals in that community identify information or products most likely to be interesting to them or relevant to their needs. These systems use similarity between user and recommenders or between items to form recommendation list for user. They do not take into consideration social trust network between entities in society to ensure that user can trust recommendations received from system. We have proposed a model where trust network exists between peer agents and recommendations are generated on basis of these trust relationships. The user agent seeks recommendations from trustworthy peers and produces aggregated order of recommendations taking degree of trust on recommenders into consideration. This paper also suggests a method to update trust on recommender based on importance of recommendation for user.