Detecting Fake Reviews Utilizing Semantic and Emotion Model

As people are spending more time to shop and view reviews on line, some reviewer write fake reviews to earn credit and to promote (demote) the sales of product and stores. Detecting fake reviews and spammers becomes more important when the spamming behavior is becoming damaging. This paper proposes three types of new features which include review density, semantic and emotion and gives the model and algorithm to construct each feature. Experiments show that the proposed model, algorithm and features are efficient in fake review detection task than traditional method based on content, reviewer info and behavior.