A Hierarchical System Of Trust In E-commerce Model

With the development of Internet, E-commercial brings huge changes to people’s shopping life. But users of E-commercial system are very confused because of the Information Overflow of E-commercial system. Recommendation systems(RSs) plays an important row on alleviating Information Overflow and improving user satisfaction. Recommendation System brings convince to users and profit to businesses. So far, one of the most successful recommendation technology is collaborative filtering which suffering from some problems like data sparseness, cold start, sensitive to malicious attacks and poor scalability. In order to alleviate or even overcome these problems, scholars in this research field try to import trust into RSs, as a result, the application of trust in RSs becomes a popular researching topic in the current academia.In this paper whose main topic is how to improve user’s satisfaction, the users of RSs are classified as ordinary users, cold start users, controversial users. For the calculation of trust degree, this paper imports trust into RSs through social networking and references knowledge of sociology and identifies trust source among people as three levels:internal trust, interactive trust and external trust. The internal trust and interactive trust are local trust and the external trust is global trust. Using empirical parameters, this article combines three levels of trust to enable this model efficacious for different types of users. During the experimental part of this research, we explore the empirical parameters of the hierarchical model basing the distributed computing model and the data set of Epinions.com which is very pupular, and then we find different parameters for different type of users. In the end, this paper compared the recommendation results of hierarchical model with the trained parameters.