Comparison of Naïve Bayes and SVM in Sentiment Analysis of ChatGPT for Learning on X and YouTube

Abstrak

The rapid development of artificial intelligence technology has encouraged users to actively express opinions on social media platforms such as X and YouTube, including discussions on the use of ChatGPT as a learning support tool. This study aims to analyze public sentiment toward the use of ChatGPT in learning contexts by comparing the performance of the Naïve Bayes and Support Vector Machine (SVM) classification methods. A total of 5,500 comments from platform X and 5,543 comments from YouTube were collected through a crawling process using relevant keywords during the period from January 2023 to December 2025. The data were preprocessed and labeled into three sentiment classes (positive, negative, and neutral) using a lexicon-based approach with the INSET Lexicon. Feature extraction was conducted using the Term Frequency–Inverse Document Frequency (TF-IDF) method, and the dataset was divided into training and testing sets with an 80:20 ratio. Model performance was evaluated using accuracy, precision, recall, and F1-score. The results show that the SVM classifier consistently outperformed the Naïve Bayes method on both platforms. On platform X, SVM achieved an accuracy of 76.67%, while Naïve Bayes obtained 74.60%. On YouTube, SVM achieved an accuracy of 73.10%, significantly higher than Naïve Bayes at 62.04%. These findings indicate that SVM is more effective for sentiment analysis of social media data related to the use of ChatGPT in learning environments