Spotting Suspicious Reviews via (Quasi-)clique Extraction

摘要: How to tell if a review is real or fake? What does the underworld of fraudulent reviewing look like? Detecting suspicious reviews has become major issue for many online services. We propose use clique-finding approach discover well-organized reviewers. From Yelp dataset with over one million reviews, we construct multiple Reviewer Similarity graphs link users that have unusually similar behavior: two reviewers are connected in graph they reviewed same set venues within few days. these graphs, our algorithms extracted large cliques and quasi-cliques, largest containing striking 11 who coordinated their activities identical ways. Among detected cliques, portion contain Scouts paid by new areas. Our work sheds light on little-known operation.