A fine grained evaluation and mining of E-commerce feedback comments

Abstract

Electronic commerce(E-commerce), is the trading of products and services through Internet. It allows the customers to enter their opinions about the product and purchase, which help new customers to analyze the product and to take appropriate decisions. But reading the entire comment and reach into a conclusion is difficult, so accurate evaluation of these comments is needed. The computers are becoming smarter day by day so the customers are now able to express their emotions along with their comments. Customer satisfaction is reflected mainly as feedback in the form of comments where emoticons are the main representative of emotions. So whenever emoticons are used, their associated sentiment dominates the sentiment conveyed by text comments. Also the comments may contain modifier words to give more strength to their opinions. This paper propose a modified algorithm to calculate a more accurate rating that considers the text polarity, meaning, strength, modifiers and sentiment of each sentence.