Author(s): Krina Patel and Moayed D Daneshyari

Abstract: According to consumer statistics, 74% of online consumers read reviews before purchasing a product. Consumers believe that the reviews are written by people who have purchased the same product. However, that is not the case sometimes. Many reviews are fake and generated by bots to enhance the brand value or demolish some product’s image by posting a negative review for it. Limited research has been done in this area and automatic detection systems show partial success in detecting fake reviews. In this project, we discuss the issue of fake reviews and methods to detect them. The project experiments with three models – Naïve Bayes, Support Vector Machine, and Random Forest for classification. By observing the results of these models, we can surely say that human eyes cannot detect computer-generated reviews as accurately as machine learning techniques can. This web application can be implemented in any e-commerce platform to train, and test based on their data, and it can provide consumer protection and increase the credibility of reviews.

DOI: 10.33545/27076636.2023.v4.i1a.83

Pages: 85-89 | Views: 176 | Downloads: 51

Download Full Article: Click Here