A Lexicon based Unsupervised Model to Evaluate Product Ratings Vs Reviews
E-Commerce has emerged as the new paradigm for purchase. E-commerce shoppers would like to look at customer review as a reliable source of information. These e-commerce websites provide features for customer to write the product reviews and scores the product from 1 to 5 or it’s commonly referred to as star rating. But sometimes it is seen that there are inconsistencies between the star ratings and the reviews. Because of that it is necessary to validate the star rating versus the reviews. There may also be cases where the customer would have given the review without rating the product. In such cases, we would also like to predict the star rating for a given review. In this work we would like to propose a framework based on text analytics using unsupervised sentiment analysis  and KNN based regression  to provide the star rating for the reviews and validate the ratings with respect to the reviews. We have taken 2000 product reviews from Amazon and applied this methodology. The results found was encouraging and we would like to apply other techniques to validate and do deeper analysis to establish the actual star rating.
Published in: Proceedings of the Seventh International Conference on Business Analytics and Intelligence, December 2019, IISc, India.