Fake Review Detection

Abstract

Online reviews are often the primary factor in a customer’s decision to purchase a product or service, and are a valuable source of information that can be used to determine public opinion on these products or services. Reviews provide feedback to the companies about their product for any kind of improvement. The huge impact of reviews on customer’s decision making motivates wrongdoers to create fake reviews to deliberately promote or demote a product. This is known as Opinion (Review) Spam, where spammers manipulate and poison reviews (i.e., making fake, untruthful, or deceptive reviews) for their profit. In order to provide right information to the customer detection of fake reviews is important. Manual detection of fake reviews is a time consuming task therefore we need an automated technique to detect the fake reviews. Natural Languages Processing(NLP) can be used to extract meaningful features from text content of reviews therefore it is possible to detect fake reviews using various machine learning techniques. In order to influence people fake review writers try to use words or topics that create an impact on readers mind. This difference in word choice pattern in fake and truthful review can be used as a method to identify fake reviews. Our work is based on this topic type differentiator to evaluate individual reviews. It is seen that fake review writers use words that are different from the truthful ones based on this an automated method is created using various machine learning techniques to segregate fake and truthful reviews. The method improves the efficiency and performance of fake review detection.