Sentiment analysis of client reviews of the Russian Agricultural Bank service and predicted rating reviews

Abstract

The paper provides a sentimental analysis of reviews of the Russian Agriculture Bank service. An analysis of the sentiment of a text is a class of methods for the content analysis of texts in computer linguistics for the automated detection of emotional assessment-sentiment. As methods, we use XGBoost, neural networks. In work, we determine the tone of reviews with a rating of three, since we can not always attribute reviews to good or bad; we also determine the sentiment of reviews without a score. As a result, it was determined that the main array of reviews about the bank's service has a negative sentiment.