<![CDATA[This study aims to predict emotions in Indonesian text using the IndoBERT model. Emotions play an essential role in human communication and have a significant impact on sentiment analysis and natural language processing. In Indonesia, the lack of optimized datasets and models for emotion analysis in the Indonesian language poses a major challenge. This research utilizes IndoBERT, a BERT-based model specifically trained for Indonesian, to predict six categories of emotions: anger, sadness, happiness, love, fear, and disgust. The research methodology includes data collection from social media X, data preprocessing, emotion labeling, model training, and performance evaluation using metrics such as accuracy, precision, recall, and F1-score. Results show an overall model accuracy of 73%, with strong performance in recognizing emotions like "disgust" and "fear," although there are misclassifications in distinguishing similar emotions like "happiness" and "love." These findings indicate that IndoBERT has significant potential for emotion prediction in the Indonesian language and provides a foundation for developing more culturally relevant NLP technologies for Indonesia.]]>
Copyrights © 2025
