Sentiment Analysis of Kaspi Product Reviews

Customer reviews are the key to the successful functionality of the companies. A producer will gain the true result of his products from the customer feedback. What should the seller do to achieve good quality for his product? Just by looking at the stars and reading the usual text, the seller will not be able to improve their product well enough, because different people can understand the same text in different ways. That’s why the seller can improve the wrong part of their product and make their product worse. Sentimental analysis can be applied here. He can make necessary changes to his product according to the feedback. However, it is common to see when the most clients may inadequately overestimate or underestimate the mark. To avoid confusion, it is better to analyze the text. This paper focuses on the sentiment feedback analysis and finding the relationship between reviews and ratings based on data collected from the “Kaspi.kz” marketplace. These data streams are cleaned, analyzed and reviews are got through opinion mining. After we performed data augmentation. We used several algorithms on both data to find the best one.