<![CDATA[The emergence of social media as a digital public sphere has opened significant opportunities to capture public opinion directly, including on social programs such as the Koperasi Desa Merah Putih. This program has attracted considerable attention on Twitter; however, public perception remains mixed, ranging from support and skepticism to criticism. This study aims to understand public sentiment through text analysis using two classification algorithms: Naive Bayes and Support Vector Machine (SVM). The results show that Naive Bayes performs better in balancing predictions, achieving an AUC of 0.71 and an F1-score of 0.62. In contrast, SVM, despite a slightly higher accuracy (0.69), was only effective in identifying the neutral class. This tendency indicates that Naive Bayes is more capable of capturing opinion variations in short and unstructured texts such as tweets. Furthermore, the sentiment distribution reveals a dominance of neutral opinions, suggesting the public's inclination to share factual information without explicit emotional expression.]]>
Copyrights © 2025
