<![CDATA[Social media has become a means for the public to express their opinions on various events, including the floods in Aceh. This study aims to analyze public sentiment and compare the performance of the Naïve Bayes, Random Forest, and Support Vector Machine (SVM) algorithms in classifying sentiment on the X (Twitter) platform. In addition, this study also evaluates the effect of applying the Synthetic Minority Over-sampling Technique (SMOTE) on improving model performance.The dataset was collected using a crawling method utilizing Twitter Harvest on the X (Twitter) platform during the period from November 18, 2025, to January 5, 2026. The data collection process yielded 1,971 Indonesian-language data points, which after preprocessing stages such as text cleaning, stemming, and duplicate removal resulted in 1,874 data points. The dataset was then divided into 80% training data (1,499 data points) and 20% test data (375 data points). The analysis results show that the majority of public opinion has a positive sentiment of 77.9% (1460 data), while negative sentiment is 22.1% (414 data). The model evaluation results show that the application of SMOTE can improve the performance of the three algorithms. The algorithm with the best performance is Support Vector Machine (SVM) with an accuracy value of 83% and an F1-score of 75% after the application of SMOTE. Based on the results of the study, the SMOTE technique has been proven to help improve the model's ability to recognize minority classes, resulting in better classification performance.]]>
