Fake Product Review Monitoring System

As e-commerce and internet buying have grown in popularity, it is impossible to underestimate the significance of genuine product reviews. False evaluations which are meant to mislead prospective clients provide serious difficulties for both customers and companies. False reviews can be written by both the buyer to mislead the other customers and the seller to degrade the product quality and sale of the competitor. This study presents a complete system for monitoring fraudulent product reviews that makes use of sentiment analysis, content similarity and review deviation techniques by assuming some threshold values. Along with this the best model is been saved for the future prediction on the new unseen reviews. Also the system uses three potent machine learning classifiers: Random forest, Logistic regression and support vector machine and after the comparison it has been observed that Random forest best performs out of the other. It is observed that all three methods contribute in monitoring the fake reviews patterns.