Fake Review Detection in E-Commerce: A Machine Learning Approach Using Natural Language Processing

Techniques using Natural Language Processing (NLP) to identify and remove bogus reviews are highly useful in the e-commerce sector. In this work, we employ two different machine learning models — Naive Bayes and XGBoost — to classify whether a given dataset is authentic or not. Fake reviews undermine user trust, making this a critical challenge for e-commerce companies to address.By building this model, it can help the owners of the websites and the application developers identify these fake reviews and therefore can preserve the consumer's confidence in online shopping. In this paper proposed Naive Bayes and XGBoost classifiers, the model very efficiently gives out the results of the probabilities of the likelihood of the review to be fake or not. Although trained, only Amazon and Yelp datasets achieved an impressively high accuracy score using XGBoost. Its scalability suggests that using this model on larger datasets could further enhance its accuracy and adaptability in identifying fake reviews.