System of Automatic Determination of Ukrainian Text Tone

Abstract

In the work, the system of determination of the emotional tone of Ukrainian texts based on dictionaries and rules is proposed. The developed software downloads text information in various formats and carries out tokenization and lemmatization procedures using the Python Tokenize UK and pymorphy2 libraries. The obtained array of words in the basic grammatical form was analyzed using tonal dictionaries of the Ukrainian language. A dictionary of synonyms was used to expand the vocabulary. To increase the validity of sentiment analysis, coefficients were used that take into account the various emotional loads of words of different speech parts and their dissimilar impact on the overall assessment of the text tone as well as the action of intensifying or softening adverbs. Using the means of fuzzy modeling for sentiment analysis of texts makes it possible to take into account subjective factors of the expression of human emotions and for the contribution of all emotional categories in the final evaluation of the text tone. A quantitative value of the text tone has been obtained by the method of the center of gravity for one-element sets. Based on the analysis of Ukrainian-language texts, it was found that emotionally significant words have a greater impact on the text tone value of short messages.