Analisis Sentimen Terkait Isu Penyebaran Nyamuk Wolbachia di Indonesia Pada Platform YouTube Menggunakan Naive Bayes

The spread of Wolbachia mosquitoes in Indonesia has become a topic of debate in society, with various opinions emerging regarding its effectiveness and impact. However, the variation in sentiment has not been analyzed in a structured and measurable way to determine the polarization of public sentiment on YouTube. Therefore, this study aims to analyze public perception of the Wolbachia program using a Multinomial Naïve Bayes sentiment analysis method on YouTube comments. The research method includes text data preprocessing, word weighting using TF-IDF (Term Frequency-Inverse Document Frequency), and the application of three classification approaches: Multinomial Naïve Bayes, Multinomial Naïve Bayes with SMOTE, and Multinomial Naïve Bayes with SMOTE and GridSearchCV for hyperparameter optimization. The results show that the majority of public sentiment towards the Wolbachia program is negative (67%), while 23% is positive and 10% is neutral. From the model evaluation, the approach without data balancing (Model 1) provides the best accuracy of 70%, although it has weaknesses in recognizing neutral and positive sentiments. Meanwhile, the approach using SMOTE and GridSearchCV (Model 3) produced an accuracy of 69% with a better balance between precision and recall. Word cloud analysis shows that negative opinions are often associated with conspiracy theories, distrust of the government, and concerns about environmental impacts. The conclusion of the study shows that public perception of Wolbachia is still influenced by a lack of education and transparency of information. Therefore, there is a need for more effective socialization and communication to reduce misinformation and increase public acceptance of this program.