The trust in trust-based recommender systems

the growth of e-commerce sites has posed a new challenge of information overload. It refers to the problem that caused by presence of too much information and resulted to difficult understanding and making decisions. Recommender systems help users to find the items of their interest from huge databases. Although collaborative filtering is the most successful technique for recommender systems, it suffers from several inherent issues such as data sparsity, cold start users, cold start items, low accuracy, and malicious attacks. To solve such issues, trust-based approaches have been proposed. These approaches use trustworthiness as a new factor to improve accuracy of recommendation especially in case of sparsity and cold-start users. This paper describes process of trust-based recommendation and carefully discusses about characteristics of trust, such as trust measurement metrics, visibility, value types, properties, dynamicity, propagation, etc. We also reviews the most important trust-based approaches

The trust in trust-based recommender systems Keywords:

component , Trust-based approaches , Recommender systems , Characteristics of trust