Generating Trust Based Recommendations for Multiple Independent Domains

Recommender systems generate responses and suggest items in the required domain. This paper proposes a domain independent trust based recommender system where personalized recommendations can be generated for any domain known to the recommenders. In this recommender system, there exists a web of trust which is formed on the basis of trust among agents in application domain. Here each user captures his likes and tastes for various domains in the form of upper or lower bound of each attribute of items belonging to domain under consideration and stores this information in his profile which then forwarded to his recommender agents. The recommender agents pass on only those recommendations to the user agent that matches its tastes leading to the personalization of recommendations. Experiments were conducted on a real dataset and results illustrate the proficiency and effectiveness of the proposed method.