Securing online reputation systems through temporal, trust and user correlation analysis

Online reputation systems play a significant role in decision support during internet mediated service provision. With the fast development of online reputation systems in various social networks, operations against such systems are growing rapidly. It has been considered as an important but most difficult problem. In this paper, we propose a defense solution TTAUCA [Temporal, Trust And User Correlation Analysis], which secure the reputation systems from a novel angle, that is used to integrate time domain anomaly detection, Dempster-Shafer Theory based trust computation and user correlation analysis. Real user attack data collected from a cyber competition is used to construct the testing data set. Compared with previous scheme TATA [Temporal And Trust Analysis], TTAUCA achieves a better performance in terms of detecting products under attacks, identifying malicious users who insert illegitimate ratings, and recovering reputation scores of victim products.