Predict Star Rating from Review text-Yelp

This research paper was completed in fulfillment of requirements for the Search course as a part of the Master’s in Data Science program at Indiana University-Bloomington in May 2019. Reviews play an important role to understand information about event, place or any commercial product. These reviews are captured simply on the numeric scale or it can be an elaborated details text review. In this study, we analyze if we can summarize the big textual review by predicting stars using information retrieval and machine learning. We also perform sentiment analysis to understand if the sentiments in the review are in accordance with the star rating provided. Our findings show that star rating can be predicted to some extent based on the review comments. The sentiment present in review some time doesn’t match with the star rating provided for the review. This model can be enhanced further to add more features to improve the accuracy which will be beneficial to summarize the review comment to the star rating and save people time to read through the entire review description