Research Progress of Trust Evaluation

Trust evaluation is one of the most important issues in trust related research. How to evaluate the trust between two users is the main problem faced by many current recommendation systems and trust research. Currently in many applications, such as movie recommendation, spam detection, and online borrowing, evaluating trust among users in a trust social network (TSN) is a key issue. Therefore, this paper introduces the development process of trust evaluation in two aspects. The first is trust evaluation under different factors, such as user information and evidence. The second is trust evaluation based on different methods, such as neural networks and collaborative filtering methods. In the future, more factors can be combined with neural networks and reinforcement learning for trust assessment. For user privacy protection, blockchain technology can be combined to better encrypt user information, making the results more accurate and close to reality, and apply to more recommendation systems.