THEORETICAL ISSUES OF SENTIMENT ANALYSIS IN THE UZBEK LANGUAGE

In this article, general theoretical issues of sentiment analysis of Uzbek language texts are analyzed. Linguistic knowledge database and corpus are necessary for automatic analysis. Sentiment analysis can be neutral, positive or negative depending on the attitude of the commenters. According to the analysis of the studies in this sphere, there are four components of the expressed attitudes: the subject of the attitude, the object of the attitude, the aspect of the attitude, the type of attitude. Accordingly, the corpus and thesaurus are used as research objects in the sentiment analysis of Uzbek texts. Accordingly, the distinction of people’s attitudes relied on various attributes such as feelings, mood, social, emotional attitude, and personal behavior cited with examples in the article. It is claimed that to apply scientific results in this area could be applicable such as WordNet-Affect, SentiWordNet, SenticNet, MPQA Opinion Corpus, RuSentiLeks software, which have been tested in the world experience of creating the Uzbek language sentiment analysis program.