Semantic Knowledge Base for the Emotional Coloring Analysis of Kazakh Texts

Lexical resources such as dictionaries and thesauri are crucial for solving various tasks in the field of natural language processing (NLP). This paper presents the development and application of a semantic knowledge base specifically designed for sentiment analysis of Kazakh texts. Given the low-resource nature of the Kazakh language, there has been a significant gap in available lexical resources with annotated sentiment. Our work addresses this gap by creating a comprehensive semantic base consisting of over 16,000 words and phrases, annotated on a 5-point sentiment scale ranging from -2 (very negative) to 2 (very positive). This base includes various parts of speech such as adjectives, nouns, verbs, adverbs, interjections, and emoticons, and considers contextual variability in word sentiment. We also developed and implemented a hybrid sentiment analysis model that leverages this semantic base to accurately determine the sentiment orientation of Kazakh texts at both the sentence and document levels. The effectiveness of this model is demonstrated through its application in various domains, including social media monitoring, marketing, healthcare, and education. Our approach opens new opportunities for research and practical applications in the field of natural language processing for low-resource languages.