Sentiment Analysis of Twitter Data on Indonesia’s Cabinet Using Naïve Bayes and Support Vector Machine Algorithms

Twitter has become a widely used platform for information dissemination among internet users and it serves as a valuable data source for sentiment analysis and decision-making. In this context, sentiment analysis is used to automatically categorize user tweets into positive or negative opinions. The Indonesia Maju Cabinet, the current administration under President Joko Widodo has emerging various public opinions regarding their performance and responsibilities. Sentiment analysis provides a method to categorize public opinions on social media. This study uses a dataset collected through a crawling process on Twitter with the keyword "Menteri Jokowi" (Jokowi's Ministers). The obtained data was then analyzed using two algorithms: Naïve Bayes Classifier (NBC) and Support Vector Machine (SVM), to compare their cross-validation results. The analysis results show that the Naïve Bayes Classifier algorithm achieved 91.70% accuracy, 91.69% recall, and 91.69% precision. Meanwhile, the SVM algorithm achieved 96.77% accuracy, 96.71% recall, and 96.71% precision. The difference in accuracy is due to NBC’s tendency to misclassify neutral tweets as positive, whereas SVM, despite optimizing class separation, struggled with detecting sarcasm and subtle sentiment shifts, sometimes misclassifying negative tweets as neutral. Based on these results, it can be concluded that both algorithms can be effectively used for classifying opinions about ministers through sentiment analysis, although SVM demonstrates higher accuracy.