A Unified Model for Unsupervised Opinion Spamming Detection Incorporating Text Generality

Yinqing Xu, Bei Shi, Wentao Tian, Wai Lam

Abstract: Many existing methods on review spam detection considering text content merely utilize simple text features such as content similarity. We explore a novel idea of exploiting text generality for improving spam detection. Besides, apart from the task of review spam detection, although there have also been some works on identifying the review spammers (users) and the manipulated offerings (items), no previous works have attempted to solve these three tasks in a unified model. We have proposed a unified probabilistic graphical model to detect the suspicious review spams, the review spammers and the manipulated offerings in an unsupervised manner. Experimental results on three review corpora including Amazon, Yelp and Trip Advisor have demonstrated the superiority of our proposed model compared with the state-of-the-art models.