Stages of Sentiment Analysis of Uzbek Texts Using the ABSA Method

Aspect-Based Sentiment Analysis (ABSA) is one of the most actively studied directions in Natural Language Processing (NLP) today, as it enables a deeper understanding of user opinions. The volume of user-generated content on social media is vast, and manually analyzing such data requires substantial time and effort. Traditional sentiment analysis provides a general sentiment score for the entire text, which may overlook some important details and specific opinions. In contrast, ABSA identifies and evaluates each aspect mentioned in the text individually, allowing for a more fine-grained sentiment interpretation. This paper presents the global significance of ABSA, its stages of development, and the theoretical foundations for its application to texts written in the Uzbek language. Considering that ABSA research in Uzbek is still in its early stages, the focus of this study is primarily theoretical. It aims to systematically describe the ABSA process adapted to the linguistic characteristics of the Uzbek language. Each stage of the ABSA pipeline directly influences the overall performance and accuracy of the system. In particular, the complexity of Uzbek morphological structures, the language’s free word order, and the frequent use of idioms and metaphors in context require customized analytical approaches. In such cases, it becomes essential to develop models tailored to the unique linguistic features of Uzbek.

Ҳали таржима қилинмаган