<![CDATA[Danantara a state-owned investment management institution, has become a topic of widespread public discussion, particularly on social media platform X, where diverse public opinions are expressed. This study aims to evaluate public sentiment toward Danantara through sentiment analysis using machine learning techniques. The dataset consists of 10,108 tweets, of which 9,790 tweets remained after the preprocessing stage and were ready for analysis. The methodology involves word weighting using Term Frequency-Inverse Document Frequency (TF-IDF) and the implementation of two classification algorithms: Naïve Bayes and Support Vector Machine (SVM). To address the class imbalance in sentiment data, the Synthetic Minority Over-sampling Technique (SMOTE) was applied. Model performance was evaluated using metrics such as accuracy, precision, recall, and F1-score. Initial results show that before applying SMOTE, the Naïve Bayes algorithm achieved an accuracy of 64%, while SVM performed better with an accuracy of 80%. After applying SMOTE, Naïve Bayes accuracy improved to 72%, and SVM increased significantly to 89%. These results indicate that SMOTE is effective in handling data imbalance and enhancing classification performance. Overall, this study provides a clearer picture of public opinion toward Danantara and demonstrates that the combination of preprocessing, TF-IDF, machine learning algorithms, and data balancing techniques can produce more accurate sentiment analysis.]]>
