Research on Spam Review Detection: A Survey

With the vigorous development of e-commerce, enterprises and individuals generally use network platforms to obtain information and conduct commodity transactions, providing convenience, speed and efficiency to the society and people, but also accompanied by endless fake transactions, deceptive spam reviews, and credit fraud. Online stores improve their system rankings through deceptive spam reviews and other methods to obtain profits. To solve this problem, spam review detection becomes one of the key technologies. This paper summarizes the existing research on spam review and spammer detection, and classifies spam review and spammer detection technologies to methods based on review content features, reviewer behavior characteristics, reviewer graphs and deep learning. Among them, deep learning-based methods for detecting spam reviews and spammers mainly focus on the cold start problem and deep representation of reviews and reviewers. Finally, the paper summarizes the research problems and future directions.