<![CDATA[Corruption remains a major public issue in Indonesia, including the Pertamina corruption case, which has triggered diverse reactions across social media. This study aims to analyze public sentiment toward the Pertamina corruption issue using the Naïve Bayes algorithm as a text classification method. The research process includes data collection, text preprocessing consisting of slang normalization, stopword removal, and stemming, term weighting using Term Frequency-Inverse Document Frequency (TF-IDF), sentiment classification using Naïve Bayes, and model evaluation using a confusion matrix. The dataset used in this study consists of public comments in Indonesian that were processed and classified into negative, neutral, and positive sentiment categories. The evaluation results show that the model achieved an accuracy of 0.7592 with a macro F1-score of 0.4642. The negative sentiment class produced the best performance compared with the neutral and positive classes. Dominant negative terms such as corruption, law, corruptor, and death indicate that public opinion tends to respond negatively to the case. These findings provide a clearer picture of public perception regarding the Pertamina corruption issue and may serve as a reference for stakeholders in understanding public opinion expressed on social media.]]>
Copyrights © 2026
