A Novel Trust Model Based on Recommendation for E-commere

With the enormous growth of the Internet and e-commerce, online trust has become an increasingly important issue. The overall goal of this research is to develop a support way for computational trust and reputation models based on recommendation to facilitate reliable services in online environments. This paper presents a novel recommendation-based trust model, which utilizes the theory of collaborative filtering and ant colony optimization, and includes two indexes of trust scenario similarity and propensity to trust similarity. Compared with existing trust model, this model has three contributions. First, based on the ant colony algorithm, the optimal trust path can be searched effectively. Secondly, it addresses the problem of subjectivity in path weight initialization and improves computing efficiency. Finally, comparing the model with Beth model based on eBay´s data, the result shows the feasibility and effectiveness of this approach.