Fake Review Detection using BERT Transfer Learning Algorithm

Abstract

Abstract: In this research, our main focus is on tackling the issue of fake online reviews, which has become increasingly prevalent due to the vast amount of content generated by users on various online platforms. These reviews have a big impact on what people decide to buy, emphasizing how important it is for them to be real and trustworthy. However, the rise in fake reviews has cast a shadow of doubt over the credibility of online platforms, making it imperative to find effective ways to discern the real from the fake. To address this challenge, we propose a strong model that utilizes the capabilities of advanced deep learning algorithms—BERT (Bidirectional Encoder Representations from Transformers) and Bidirectional Long Short-Term Memory Networks (BiLSTMs)—to accurately identify and filter out fake reviews. BERT excels in understanding the nuances of language and BiLSTMs effectively capture the order of words and phrases. Our plan involves training and validating this model using datasets sourced from reputable platform like Kaggle. By amalgamating these powerful algorithms and datasets, we aim to significantly enhance the accuracy and credibility of our fake review detection system. Ultimately, our goal is to restore consumer trust and empower individuals to make more informed choices in the expansive realm of online consumerism.