DOI: 10.21276/sjet

Pages: 329-334

An excellent source of collecting the reviews on specific product is various online shopping sites where people share their reviews on products and their shopping experience. People may come through the wrong opinions known as review spam. Therefore, for this it is essential to detect it by some means. In this paper, presents methods for detection of spam users using feature extraction and discretization, in combination with EM algorithm. Our framework can detect multiple spammers by knowing only small set of spammer sets. Proposed method effectively selects relevant features and builds features set to identify the spammers. In this paper, we have blocked the users with fake id or who are predicted as spammer.