Research On Recommendation Algorithm Based On Trust And Distrust Network

With the continuous development of information technology and electronic data constantly increasing, people have gradually entered an era of big data. A sharp increase in electronic data, not only brought unprecedented convenience to people’s lives, but also brought unprecedented trouble. On the one hand, the electronic information people can be obtained sharply increased. On the other hand, the amount of information that people can understand within the unit of time has no change. For this conflict, personalized recommendation system came into being. The main purpose of personalized recommendation system is, by some means, to filter out unwanted or usefulness of the information, and to retain the most useful information for this user. Use the recommended system, users just need to deal with a small amount of information, to get most of the benefit which need to traverse all the information.In recent years, the rise of social networks on people’s lives had an impact on all aspects. Various types of computer systems have gradually introduced social networking features to mobilize the enthusiasm of the user, thereby improving the user’s activity. Because of these social behavior, it has a profound impact on the behavior of users in the system. Therefore, by studying these social behavior, the system can predict the user’s behavior. This is also a hot issue in recent years. Among the many studies on the influence of social behavior for the system, the study of the influence of trust behavior for the system got more extensive attention. The basic idea of it is that people are more inclined to believe that they trust people’s views. By fusion the research of trust network to the study of recommendation system, it can further improve the recommendation effect of recommendation systems. This aspect of the research has been some fruitful research results, of course, there are also some disadvantages. This article is based on this aspect to expand the study.The main work of this article is summarized as follows:1、 In this paper, we first summarize the basic methods and the latest research progress of recommendation system. We summarize the strengths and weaknesses of their predecessors, which laid the foundation for improved algorithms below.2、 Then, we summarized the knowledge of trusted relationship network and proposed an improved scheme. This scheme’s first work is to expand the trust network, by using trust transfer and trust aggregation. Then, from both directions of trust and untrusted, we re-evaluation of the degree of trust between users, so as to establish a trust which is more in line with the characteristics of the system. The experimental results show that the trust relationship network built with new algorithms reflect the characteristics of the trust relationship between users of the system better, which can also improve the recommendation algorithm performance.3、 On the basis of the proposed new trust relationships network, we designed and implemented a new recommendation algorithm which combines the advantages of the new trust relationships network and collaborative filtering algorithm. This algorithm is very easy to implement and the results are easy to interpret. There are many relevant comparative study, so that we can easily verify the effect of trust relationship networks. Experiments show that such new novel algorithms improved the recommendation effect of the recommendation algorithm from multiple angles.