Exploring Machine Learning and Deep Learning Techniques for Fake Review Detection: A Comprehensive Literature Review

Abstract

Over the last few years, the impact of reviews on ecommerce industries and the folk who rely on the online reviews have increased. Most of the online shoppers depends on the reviews to make their purchasing decisions. Moreover, genuine reviews can assist the businesses acquire higher sales. Implementing effective systems can ensure the reliability of reviews and create trustworthiness on online platforms. Deceptive reviews can deteriorate the integrity of online feedback system. Numerous studies have explored this domain employing Machine Learning, Deep Learning, NLP methodologies etc., in the last decade. In this paper, through the survey we address the challenges which the current systems encounter where they lack adequate capabilities in detection and removal of false reviews. Diverse methods are applied on datasets like Amazon, Yelp etc., to obtain organized information and to improve the performance of the employed systems in classifying reviews as fake or genuine. Furthermore, this paper provides details of each method, their accuracies and also the future directions in this area. There is a pressing need of systems that can effectively address the issue of fake reviews as the existence of these can mislead customers leading to the decline in their preference for ecommerce.