Detection of Fake Reviews on Online Reviews Platforms using Deep Learning Architectures

Abstract

Fake internet reviews are on the rise, and consumer advocacy organizations and regulatory agencies are concerned. Identifying fake reviews with accurate results, on the other hand, is a challenging task in marketing and Computer Science. Traditional text classification approaches use a bag-of-words model to represent text, which causes sparsity, and word representation learned from neural networks, which has limited ability to handle unknown words. In this paper, we implement the detection of fake online reviews using various deep learning techniques and analyse the accuracy of each technique being applied. Techniques being used are Bidirectional LSTM with GLoVe 50D, CNN-LSTM, etc. It includes improvement in the attention layer to increase testing accuracy.