<![CDATA[TikTok Shop has become one of the largest social commerce platforms in Indonesia; however, user perceptions of its shopping features have not been systematically studied. Star rating assessments alone are insufficient to deeply understand user satisfaction and complaints, necessitating a more comprehensive text analysis approach. This study aimed to analyze the sentiment of TikTok Shop user reviews on Google Play Store using IndoBERT, with a contribution of a keyword-based filtering approach to extract shopping-related reviews from general TikTok application reviews. Review data were collected through automated extraction from Google Play Store, filtered using shopping-related keywords, and preprocessed before classification using IndoBERT. Evaluation was conducted using a confusion matrix with accuracy, precision, recall, and F1-score metrics. Results showed a dominance of positive sentiment, with positive reviews centered on shopping convenience and live shopping satisfaction, while negative reviews were dominated by technical complaints and delivery issues. The model achieved adequate performance on positive and negative classes, although the neutral class was undetected due to class imbalance. This study demonstrates that IndoBERT effectively analyzes sentiment in Indonesian-language social commerce reviews, and the findings serve as empirical evidence for platform developers and digital business actors in formulating service improvement strategies.]]>
Copyrights © 2026
