<![CDATA[Social media X provides many public responses to the Free Nutritious Meal Program (MBG), including support, questions, criticism, and neutral information. This study processes 6,000 tweets related to MBG to identify the direction of public opinion using a machine learning approach. The research flow consists of sentiment labeling, text cleaning, TF-IDF weighting, and model testing using Naive Bayes, Random Forest, and Support Vector Machine. The sentiment distribution shows 3,087 positive tweets, 2,213 neutral tweets, and 700 negative tweets. Model testing shows that Random Forest produced the strongest result with 94.75% accuracy, 95.66% precision, 94.75% recall, and 94.93% F1-score. These findings indicate that Random Forest is more suitable for recognizing sentiment patterns in the MBG tweet dataset than the other two models. The study also presents the analysis through a web-based system containing dashboard, dataset import, sentiment data, preprocessing, training, evaluation, and new opinion classification features.]]>
Copyrights © 2026
