A Strong And Weak Ties Feedback-Based Trust Model In Multimedia Social Networks

The multimedia social network, a combination of the multimedia sharing technology and social network, has prominent social and diffused features. Due to its centerlessness and lack of regulation, the multimedia social network has some serious network environment problems, such as spread of negative digital contents and serious data redundancy. Therefore, how to establish a trust management mechanism in the multimedia social network, which could find user’s malicious collusion behavior, limit the spread of bad information in the social network, keep the network running normally and improve its security, has become an important research direction in the field of network security. On the foundation of the research on the security threats in multimedia social networks and the analysis of classical trust models, this paper makes a deep study on the trust management of multimedia social networks. The main works are as follows:1) This paper proposes a strong and weak ties feedback-based trust model in Multimedia social networks. This model consists of a partition scheme of strong and weak ties and a strong and weak ties feedback-based trust computing model. According to Weak Ties Theory and Three Degrees of Influence in the Sociology, the users are divided by this partition scheme in multimedia social networks. This partition scheme has two steps:individual partition and global partition. In the individual partition, the similarity and the reputation value of target users are main factors to divide its neighbors into weak ties and strong ties. The global partition, after the individual partition, will divide all users of the network into two types, according the Three Degrees of Influence. On the basis of the partition of strong and weak ties, the trust computing model gets the feedback value from these two types of users. Weak ties need to compute content evaluation value and strong ties need to compute the content evaluation value and behavior evaluation value. Based on the above-mentioned values, the trust computing model uses Lyapunov Central Limit Theorem and Bayesian Estimation to calculate the period evaluation value in this trust period. Finally, using the damped window mechanism and combining with the period evaluation values in the recent trust periods, the user’s reputation value in the current trust period is calculated. The simulation experiment shows that the model is reliable in design, valid in network transmission, and it can accurately and effectively reflect the user’s trust, and effectively resist malicious feedbacks and collusion attacks.2) This paper also proposes an information dissemination model in the multimedia social networks based on the trust management. First, according to user’s operations of multimedia digital content in multimedia social networks, the state transition diagram of four operations is designed. Then, we establish the information dissemination model on the foundation of SEIR epidemic model. Combining with the reputation value given in the strong and weak ties feedback-based trust model and the trend analysis of trust and transmission probability in Sociology, the influence of the trust management scheme on the information transmission in multimedia social network is analyzed. Finally, the experiment simulation shows that the trust management can enable positive digital data to spread rapidly and widely and limit the dissemination of negative contents.