A trust-based recommender system for collaborative networks

Collaborative networks are composed by people from all over and all kind of background that meet to exchange common experience. The great challenge lies in finding significant, reliable information in view of the swelling volume of information output. Recommender systems emerged as a solution to reduce the problem of information overload. Collaborative filtering (CF) technique in particular automates the recommendation process by building on users' opinions of items in a community. However CF has limitations named cold start and first rater. The goal of this paper is to present a model using trust to minimize those problems.