Turning the Tide: Curbing Deceptive Yelp Behaviors.

摘要: The popularity and influence of reviews, make sites like Yelp ideal targets for malicious behaviors. We present Marco, a novel system that exploits the unique combination social, spatial temporal signals gleaned from Yelp, to detect venues whose ratings are impacted by fraudulent reviews. Marco increases cost complexity attacks, imposing tradeoff on fraudsters, between their ability impact venue remain undetected. contribute new dataset community, which consists both ground truth gold standard data. show significantly outperforms state-of-the-art approaches, achieving 94% accuracy in classifying reviews as or genuine, 95.8% deceptive legitimate. successfully flagged 244 our large with 7,435 venues, 270,121 195,417 users. Among San Francisco car repair moving companies we analyzed, almost 10% exhibit