SENTIMENT ANALYSIS OF YOUTUBE COMMENTS ON THE VINA CIREBON CASE USING GBM AND LOGISTIC REGRESSION

Abstract

This research aims to analyze public sentiment regarding the Vina Cirebon case on YouTube by using Gradient Boosting Machine (GBM) and Logistic Regression methods. The study leverages comments from YouTube to evaluate public opinion through precision, recall, F1-score, and accuracy metrics derived from confusion matrices. The results indicate that Logistic Regression outperforms GBM in terms of accuracy (88.29% vs. 81.48%), precision (0.854 vs. 0.739), recall (0.602 vs. 0.352), and F1-score (0.706 vs. 0.476). This suggests that Logistic Regression is more effective in capturing the nuances of public sentiment on this issue. The analysis highlights the predominance of negative sentiments, reflecting widespread public disapproval and concern. This study provides valuable insights into the public's reaction to the Vina Cirebon case, demonstrating the efficacy of sentiment analysis in understanding social issues through social media data.