Spam Consumer’s Reviews Detection for E-Commerce Website using Linguistic Approach in Deep Learning

Abstract

Detecting spam reviews is an urgent need in today's society, so that consumers can avoid being duped by internet retailers. Many websites allow users to publish reviews, which opens the door to the publication of false or misleading information, such as reviews that have been paid for. To the general public, these fabricated evaluations can cause confusion as to whether or not they should believe the review. Spam review detection has been solved using prominent Machine and Deep learning techniques. Current research is mostly focused on approaches that require labelled data, which are insufficient for online review because they represent the majority of the currently conducted research on supervised learning. The goal of this research is to expose any misleading text reviews that may be out there. A number of deep learning algorithms for spam review identification have been developed, Convolutional Neural Networks (CNNs), Multi-Layer Perceptron’s (MLPs), and Long Short-Term Memory (LSTM), variant of Recurrent Neural Networks are all examples of these types of neural networks (RNN). Traditional machine learning classifiers such as Nave Bayes (NB), KNN, and SVM were also employed to detect spam reviews, and we compared the performance of traditional machine learning classifiers with that of deep learning classifiers.