Exploiting and Ranking Dominating Product Features through Communal Sentiments

Abstract

The rapidly expanding e-commerce has facilitated consumers to purchase products online. Various brands and millions of products have been offered online. Varieties of customers’ reviews are available now days in internet. These reviews are important for the consumers as well as the merchants. Most of the reviews are disorganized so it generates difficulty for usefulness of information. In this paper we are proposing a product feature ranking framework, which will identify important features of products from online customer opinions, and aim to improve the usability of the different reviews. The important product features are recognized using two observations 1) the important features are mostly commented on by a large number of users 2) users reviews on the important features are greatly influence on the overall reviews on the product. We first identify product features by shallow dependency parser and determine customer’s reviews on these features via a sentiment classifier. Then we adopt develop a probabilistic feature ranking algorithm to conclude the importance of features by considering frequency and the influence of the influence of the users reviews given to each feature over their overall reviews.
DOI: 10.17762/ijritcc2321-8169.150682