Amazon’s Fake Review Detection using Support Vector Machine

Abstract

Online user data is crucial to the marketing process since it affects consumers' daily lives. False product reviews have a negative impact on the enterprise's capacity to analyse data and make decisions with confidence. Some users have a propensity todisseminate unconfirmed fake news on internet sites.Today, it is crucial to be able to recognise fake reviews.Many websites provide things for sale to consumers online. Purchasing decisions can be made based on product reviews and market demand. On the basis of reviews, consumers determine whether a product is acceptable for use or not. There will be hundreds of comments about the product, some of which may be false. We provide a mechanism to identify fake reviews of items and indicate whether they are reliable or not in order to distinguish between them. This approach for identifying false reviews describes the use of supervised machine learning. This methodology was devised in response to gaps because traditional fake review detection methods classified reviews as authentic or false using either sentiment polarity scores or categorical datasets. By taking into account both polarity ratings and classifiers for false review identification, our method contributes to closing this gap. A survey of already published articles was conducted as part of our effort. Support Vector Machine[2], a machine learning technique, used in our system produced accuracy of 80%.