PRODUCT REVIEW ANALYSIS

Abstract

Online reviews have become a major factor in people's decision-making and business decisions. A retailer that sells products to the web often asks or takes reviews from customers about the products they have purchased. As e-commerce grows and becomes popular day by day, the number of updates received from a customer about a product is growing rapidly. With a famous product, updates can be up to thousands. The growing popularity of online reviews also encourages the business of writing a fake review, which refers to paid human writers who produce fraudulent reviews to influence readers' opinions. Our project addresses this issue with construction a separator that takes the text of the review and the basic information of its reviewer as input and determines whether the update is reliable. This creates difficulties for the customer who may read and decide whether to buy the product or not. Problems also arise so that the manufacturer can follow and manage customer feedback. And the more difficulties he faces manufacturer because many other retailer sites can sell the same product at good prices and the manufacturer often produces many types of products. In this study, we aim to summarize all customer reviews of the product and compare based products in the review can be done in one place. This summary function is different from the standard text summary, because we are the only one information about that product in which customers have expressed their views and whether the ideas are correct or not negative. We summarize the review by choosing to rewrite some of the original comments, from reviews to great snap. Points as in the summary of the ancient text. Our experimental results using a review of the number of products sold online are indicative strategic efficiency.