Spam Review Detection and Removal in E-Commerce Website

Abstract

In this era, most of the people are eager to purchase products via online e-commerce sites such as Flipkart, Amazon, Paytm., etc. Before purchasing any products via online people spend some time for analyzing the available comments such as feedbacks, reviews and ratings related to the product which are posted by other customers. Based on these comments, people took a decision whether to buy that product or not. If the given comments are positive, then there is a chance to buy the product. If not, people may ignore the product. On the other hand, these reviews play a vita role in the improvement of a business. Positive reviews can add profit to a company. While negative reviews can potentially impact credibility and cause economic loss to a company. For gaining profit to a company or making a company reputation down, imposters namely spammers were encouraged to write fake reviews about a company product. These fake reviews can mislead the customers in taking decision. If the fake reviews are positive, then most of the customers trust them and buy the product there by improve a company profit rate. Where as in case of negative reviews, the customers won’t buy the product there by damage the reputation of a company. Therefore, it is most important to detect the spammer and spam content to make online purchase trustworthy. In the past few years, several approaches have been proposed to deal with the problem but there are still some millstones for detecting the spammer and spam content correctly. The proposed framework includes two methodologies namely, Rate Deviation using Threshold and User Behavioral method for identifying spammer and spam contents in an effective manner. In Rate Deviation using Threshold method, it calculates a score value against each user rating and overall rating of a product. Whereas in User Behavioral method, it calculates the user burst time.