A Survey on Machine Learning and Deep Learning Techniques for Fake Review Detection in E- Commerce Applications

Abstract

The exponential growth in the e-commerce applications have turned the online consumer reviews an essential factor in buying decisions. However, this reliance has created an incentive to rapidly increased fake or deceptive or false reviews, (i.e. opinion spam) that compromises the integrity of online platforms. This work presents a critical survey of the recent developments in fake review detection, discussing the techniques that range from traditional Supervised and Semi-Supervised machine learning algorithms to Advanced Deep Learning and Hybrid Fusion models. We investigate the performance of different algorithms such as Support Vector Machines (SVM), Convolutional Neural Networks (CNN), Naive Bayes, and aspect-based fusion models on standard datasets. Besides that, we also discuss the role of feature engineering especially with the integration of behavioral, textual, and aspect-level features. The paper concludes by pointing out the current limitations such as scarcity of labelled datasets and also suggests the future direction.