Fine-Tuning Transformers for Multi-Label Emotion Classification: Experiments on Kazakh News Texts

Abstract

This paper presents a multi-task learning approach for multi-label emotion classification in Kazakh-language news articles using transformer-based models. We constructed a labeled dataset of 600 news articles by collecting and analyzing real user reactions from the official Kazakh news platform NUR.KZ. Each article was annotated with one or more emotions based on aggregated user feedback. A pre-trained transformer model was fine-tuned to predict these emotional reactions, demonstrating effective performance in detecting multiple emotions simultaneously. Our model achieved low loss values during training and validation, indicating good convergence, and high accuracy across the emotion categories. These results highlight the effectiveness of using naturally occurring user feedback for emotion modeling in low-resource languages and suggest promising applications in other domains such as social media and public opinion analysis.

More by Gaziza Yelibayeva

• Ontology-Based Digital Learning Platform for Kazakh Language in Primary School: Curriculum-Aligned Semantic and Adaptive ApproachIngénierie des systèmes d information · 2026 · Citations: 1
• Directional Asymmetry of Domain-Adversarial Adaptation in Cross-Corpus Speech Emotion RecognitionIEEE Access · 2026
• Constituency Syntax Parsing of the Kazakh LanguageLecture notes in networks and systems · 2026
• Сomparative analysis of grading models using fuzzy logic to enhance fairness and consistency in student performance evaluationCogent Education · 2025 · Citations: 17