Instruments and Systems: Monitoring, Control, and Diagnostics Annotation << Back A Review of Natural Language Processing for Sentiment Analysis: a Systematic Approach for Sentiment Analysis in Text Data A.U. MENTSIEV, A.L. TKACHENKO, R.S. ZARIPOVA The study focuses on sentiment analysis of text data using natural language processing techniques. Sentiment analysis is an important tool for understanding and interpreting the emotional content expressed in text. This study conducted a sentiment study based on customer reviews and product ratings on the Ozon platform. The paper presented an approach based on the VADER NLTK model, which provides a simple and effective way to analyze sentiment. The study also examined Hugging Face’s state-of-the-art ROBERTA model, which is based on a transformer architecture that allows for more accurate sentiment analysis results. The results of the study confirm that both approaches are highly effective in analyzing sentiment in customer reviews on the Ozon platform. They can be used for automatic classification and sentiment analysis, which is a valuable tool for companies in understanding the opinions and preferences of their customers. Keywords: natural language processing, sentiment analysis, text data, machine learning, deep learning. DOI: 10.25791/pribor.10.2023.1445 Pp. 08-13.