APPLICATION OF MACHINE LEARNING FOR USER SENTIMENT ANALYSIS IN INFORMATION AND COMMUNICATION SYSTEMS

The article examines modern methods of applying machine learning and recommendation systems for sentiment analysis of users in information and communication environments. Social networks and digital platforms have become important sources of public opinion, generating large volumes of textual data daily. Traditional analysis methods, such as lexical approaches or classical machine learning algorithms, have limitations in detecting context, sarcasm, slang, and emotional nuances in the text. This complicates the accurate identification of user emotions and socially significant topics. In this regard, the study proposes an effective model that combines the BERT (Bidirectional Encoded Representation from Transformers) method and a recommendation algorithm for an in-depth analysis of textual data. The developed model not only classifies the emotional tone of the text but also identifies key topics that gain social resonance, allowing it to quickly adapt to dynamic changes in the information environment. The proposed approach simplifies automated public opinion monitoring, personalization of information flows, and efficient management of unstructured data. The research results confirmed the effectiveness of the developed system: the model’s accuracy progressively increased during training—from 60% at the initial stage to over 98% at the final stage for training data. For test data, the classification accuracy reached 100%, indicating a high capacity for information generalization and a low probability of error. The minimization of the loss function confirms the efficiency of the training process and the reliability of the proposed algorithm. Integrating BERT models into communication and information systems offers vast opportunities for automated text data analysis. This approach not only improves content analysis quality but also enables the rapid detection of socially relevant topics, which is particularly important for social platforms, media analysis, and digital communications. The proposed model can significantly enhance the efficiency of information flow management, especially in artificial intelligence, automated public opinion analysis, and social event monitoring. Further research may focus on expanding the model’s capabilities for multilingual content analysis, improving its adaptation to new writing styles, and enhancing the processing of short, incomplete, or informal texts. In the future, the proposed approach may be applied to the automated management of large volumes of data, contributing to the development of intelligent information services and a better understanding of social interactions in the digital environment.