<![CDATA[In the rapidly evolving landscape of digital politics, a deep understanding of public sentiment on social media has become crucial due to its significant influence on public opinion. This study aims to analyze public sentiment toward the performance of President Prabowo Subianto by utilizing data from three popular social media platforms: Twitter, TikTok, and Instagram.The classification method employed is the Naïve Bayes algorithm, implemented within the SEMMA framework, which consists of five stages: Sample, Explore, Modify, Model, and Assess. Data from each platform was collected and processed through text cleaning, TF-IDF transformation, and class balancing using the SMOTE technique. Evaluation was conducted using Stratified K-Fold Cross Validation and the F1-score metric to assess model performance.The results indicate that classification performance varies across platforms. The model achieved the highest F1-score on Twitter data (0.82), followed by Instagram (0.72), and TikTok (0.68). Overall, the model reached an average accuracy of 75.41%. These findings suggest that the linguistic characteristics and text structures of each platform significantly affect sentiment classification effectiveness.This research provides practical implications for the application of AI-based sentiment analysis in the realm of digital politics. It offers actionable insights for policymakers to monitor public opinion in real time and for system developers to design sentiment analysis algorithms that are more adaptive to the unique characteristics of each platform.]]>
Copyrights © 2025
