Sentimental Analysis of Product Rating

Abstract

The Consumer reviews are frequently regarded as crucial resources in e-commerce systems since they capture users experiences, sentiments, and purchasing intentions. All of this data may include comments and viewpoints expressed by consumers about various topics. Numerous studies have revealed that people who share similar attitudes toward things are more likely to trust one another. Because a product can be designed for multiple e-commerce websites, the reviews on Steam are more reliable as a result. In Steam, user reviews include the following information: 1) Product name, 2) Product reviewer information, including username, delivery address, products reviewed, and reviewed date, 3) Review text, 4) Review summary (Recommended/Not Recommended), and traditional manual rating 5) The percentage of other users Helpful and Not Helpful ratings, and 6) Other users comments on the product reviews. The automated five-star rating replaces the traditional manual evaluation and uses sentiment analysis of the review. Context based anomaly detection is done to remove irrelevant portions of each review. Selective-feature analysis is performed to assist vendors as well as buyers. Therefore Threshold was set using more complex criteria that involves purchase patterns. Results obtained from the proposed model is ideal for e-commerce websites. Model can be improved by addition of sarcasm detection algorithm which is quite challenging.