Deducing Synthetic Reviews Using Gradient-Boosting Technique

This research explains a comprehensive system for detecting fake, computer-generated reviews on e-commerce platforms like Amazon and Flipkart. The proposed solution integrates a robust machine learning model, a scraping mechanism for collecting reviews, and a user-friendly frontend built with React. Using a novel ensemble method combining LightGBM, CatBoost, and XGBoost, alongside natural language processing (NLP) techniques, the system achieves high accuracy in separating genuine opinion from fake ones. This study aims to improve customer trust and maintain the integrity of online marketplaces.