A Review of Different Machine Learning Techniques for Fake Review Identification

Online reviews and ratings of the e-commerce products have grown to be an essential part of the contemporary customer experience in recent years, offering insightful information about the effectiveness and quality of products and services. A lot of users make their decisions about what to buy on Internet based on the reviews and ratings of e-commerce products. Consumers frequently believe that all evaluations of a products or service are authentic. Nevertheless, a lot of companies manipulate reviews by adding fraud reviews, either by using chatbots or content writer, tricking customers into buying faulty products. Furthermore, companies could fabricate unfavorable reviews of rival products on their website to harm their brand and lower their sales. These fake reviews hurt reputation of companies in addition to mislead customers. Thus, there is a need for efficient methods to capture fraudulent reviews and improve the reliability of online reviews across variety of products and services. In fields like natural language processing (NLP), opinion mining, and machine learning, numerous researchers are experimenting with different approaches. This study offers a comprehensive survey addressing the difficulties that the existing fake review identification systems are facing. Various methods, such as CNN, RNN, RF, etc. have been used on various review datasets. In order to ensure the authenticity of online reviews, this paper finishes by noting gaps and obstacles in the field of fake review detection and highlighting the steps for new methodology.