Sentiment Analysis of TikTok User Comments on QRIS Adoption in Indonesia Using IndoBERT

This study investigates public sentiment toward the Quick Response Code Indonesian Standard (QRIS) adoption by analyzing TikTok user comments using an IndoBERT-based classifier. Leveraging a real-world dataset of 1,128 comments scraped from a representative QRIS-related TikTok video, we applied minimal preprocessing before classifying sentiments into positive, neutral, and negative categories. Model validation on a manually annotated subset of 12 sentences yielded strong precision, recall, and F1 performance for positive and negative classes. Stability and robustness tests confirmed the classifier’s consistency across variations common in Indonesian social media discourse. Our large-scale analysis reveals that 50.9 percent of comments expressed positive attitudes toward QRIS, 17.9 percent were neutral, and 31.2 percent were negative. A qualitative review identified drivers of positive sentiment, such as convenience and international acceptance, as well as barriers reflected in negative comments like transaction failures and security concerns. These insights highlight the vital role of sentiment analysis in guiding policy and user education to improve QRIS uptake. By demonstrating IndoBERT’s effectiveness in a social media context, this work advances natural language processing for Indonesian and offers actionable guidance for stakeholders seeking to accelerate digital payment adoption nationwide.