<![CDATA[This study aims to analyze public sentiment toward the EDUBLOX program on Instagram comments using the Naive Bayes and Support Vector Machine (SVM) algorithms. The research process consisted of data collection, text preprocessing, manual sentiment labeling, feature extraction using TF-IDF, model training, and performance evaluation. The labeling process was conducted manually by two independent annotators based on predefined sentiment guidelines, and annotation reliability was evaluated using Cohen’s Kappa coefficient. The obtained Cohen’s Kappa value was (\kappa = 0.7907), indicating substantial agreement and good consistency between annotators. The dataset was divided into training and testing data using an 80:20 ratio. The evaluation process used confusion matrix metrics, including accuracy, precision, recall, and F1-score, while the McNemar test was applied to determine whether the performance difference between the two models was statistically significant. The results showed that the SVM model achieved a testing accuracy of 76.67%, marginally outperforming the Naive Bayes model with a testing accuracy of 75.29%. In addition, SVM demonstrated slightly better precision, recall, and F1-score values compared to Naive Bayes. However, the McNemar test produced a p-value of 0.1366, indicating that the performance difference between the two algorithms was not statistically significant. Therefore, both models can be considered to have relatively comparable classification capabilities, although SVM showed a slight numerical advantage in sentiment classification performance on Instagram comments related to the EDUBLOX program.]]>
Copyrights © 2026
