Implementation of IndoBERT for Text-Based Emotion Classification on TikTok Application Reviews in Google Play Store

The rapid growth of social media, particularly TikTok, has generated millions of user reviews containing diverse emotional expressions. Analyzing emotions within these reviews is crucial for gaining a deeper understanding of users’ perceptions toward an application and for supporting the development of more responsive digital services. However, research on emotion classification in Indonesian-language reviews remains relatively limited. This study aims to implement the IndoBERT model for text-based emotion classification on TikTok application reviews obtained from the Google Play Store. The review data were collected using web scraping techniques, followed by text preprocessing steps such as cleaning, normalization, and tokenization. Emotion labeling was performed based on Paul Ekman’s six basic emotion categories: happiness, sadness, anger, fear, surprise, and disgust. The IndoBERT model was evaluated using accuracy, precision, recall, and F1-score metrics to measure its classification performance. The experimental results show that IndoBERT achieves strong performance in classifying user emotions, with an accuracy reaching up to 90%. These findings indicate that IndoBERT is effective for analyzing emotional patterns in Indonesian-language application reviews. Furthermore, the outcomes of this study can serve as a reference for developing more advanced sentiment and emotion analysis systems, particularly for applications that rely heavily on user-generated content. Overall, the implementation of IndoBERT demonstrates promising potential for improving automated emotion classification on digital platforms.