<![CDATA[The Free Nutritious Meal (MBG) Program is one of the Indonesian government's strategic policies that has generated various public responses on social media, particularly TikTok. This study aims to analyze public sentiment toward the MBG program using the IndoBERT Deep Learning model. The data were collected through TikTok comment scraping using the Apify platform, resulting in 14,447 raw comments. After the data cleaning process, 13,574 valid comments were obtained, and a 50% sample was selected, resulting in 6,787 comments for modeling purposes. Sentiment labeling was performed automatically using a lexicon-based approach with three sentiment categories: positive, negative, and neutral. The class imbalance problem was addressed using the Synthetic Minority Over-sampling Technique (SMOTE) on the training data prior to the IndoBERT fine-tuning process. The results showed that the IndoBERT model with SMOTE achieved an accuracy of 72.39% and a weighted F1-score of 0.73. Although SMOTE improved the representation of the minority class, it reduced the overall accuracy when compared to the model without SMOTE. Nevertheless, the model was still able to classify public sentiment toward the MBG program reasonably well. The findings of this study are expected to provide useful insights for the government in understanding public perceptions of the MBG policy through social media.]]>
Copyrights © 2026
