Product Aspect Rating of E-Commerce Feedback Comments

Abstract

With the growth of Internet blogs, social networking sites and discussion forums have gained a lot of importance. People comment on these websites to express their opinions about products and services. Opinion mining is the computational field of study of customer’s opinions, attitude and emotions towards particular aspect. Opinion mining techniques are used for mining useful information from the opinions of the users. When buying a new product buyer mostly refer the opinion of other users who have bought the product. In this paper, we proposed a product aspect rating framework, it comprises mainly of four modules viz. pre-processing, aspect identification, review classification and aspect rating module. We analyses customer reviews in E-Commerce domain. The dataset used for the framework is Amazon product review dataset, which consists of numerous customers reviews about products. The major part of our work is to assign sentiment scores to the aspects. The review dataset is firstly pre-processed after which the important aspects are identified by apriori algorithm. The dataset is then classified into positive and negative review dataset using support vector machine algorithm. Finally, the positive and negative sentiment score are calculated for all the identified aspects and presented in the form of graph which improves the usability of numerous reviews giving customers an insight about the product.