Privacy Protection: An Anti-Fraud Scheme Based on Improved Bayesian Game Model in Multimedia Communication System

Abstract

In the multimedia communication system, users are often exposed to fraud information from malicious applications (such as providing fake bids and false content), which can easily cause privacy leakage. And multimedia information pluralism make it more difficult to protect users' privacy. To solve these problems, this article proposed an anti-fraud scheme based on improved Bayesian game model. First, we designed a Bargain-bayesian game model for modeling the interaction between applications and regular users. We used the Two-round bargain method to establish payoff matrix for two players and adjust it through regular user's detection rate dynamically. Then we obtained the application's best bid in the first round of bargain by backward induction to prevent malicious applications from sending fake bids. Second, we customized a group of test users and developed another interaction model between applications and users (i.e., regular and test) as the Two-side Bayesian game model based on sliding adaptive logistic regression method, then we put three influencing factors into the payoff matrix: test user's ratio, malicious application's ratio, and regular user's ratio. Through Bayesian Nash Equilibria analysis, we obtained values of three influencing factors when malicious applications provide true content, and thus solved the problem of malicious applications providing false content. Finally, experiment results proved that the new scheme had effectively raised expected payoffs for both players and their transaction achievement rate, and lowered the probability of users being deceived by malicious applications, which successfully solved the issue of users' privacy disclosure.