<![CDATA[The Free Nutritious Meal (MBG) program faces implementation challenges regarding distribution, menu quality, and budget sustainability, sparking diverse public discourse on social media. This study analyzes public sentiment toward the MBG program using an Aspect-Opinion-Qualifier Extraction (ASOQE) approach based on policy taxonomy. The dataset was obtained from X (formerly Twitter) via web scraping and processed through standardized text preprocessing. Automatic annotation used a lexicon-based BIO labeling approach to generate a silver-standard dataset. The classification model was trained using an IndoBERT-BiLSTM architecture to identify contextual aspects and opinions. Inference results were mapped into five sentiment classes and five policy dimensions: nutritional quality, implementation, social impact, policy, and effectiveness. Evaluation showed excellent performance, with F1-scores exceeding 0.98. Findings reveal that social impact and implementation dimensions dominate public discourse, showing significantly positive sentiment. This research demonstrates the potential of Aspect-Based Sentiment Analysis as a data-driven tool for comprehensive public policy evaluation.]]>
Copyrights © 2026
