Development of Logical Methods for Extracting Emotional Assessments from Natural Language Texts

The article is devoted to the problem of extracting emotional assessments from natural language texts. The article analyzes existing methods of sentiment analysis and explores modern methods of natural language processing to determine the sentiment of Uzbek texts. Existing approaches to the study and description of emotions are explored. Particular attention is paid to emotional assessments of situations presented in natural language texts. Logical-semantic methods and technologies based on deep machine learning methods are considered. Formalization of emotional assessments contained in natural language texts is carried out using the theory of partial models. The task of classification is to assign a class to objects based on their characteristics. In the case of the emotion model, the emotion texts are object-class pairs. Since the number of emotions is not binary, the classification is not binary. To solve the classification problem, a training data set consisting of labeled text with emotions is required.

Перевод пока недоступен