Analisis Sentimen Tentang Penggunaan Galon Bebas BPA di Indonesia Menggunakan Algoritma Support Vector Machine

This research employs the Support Vector Machine algorithm to classify sentiment in comments on the X, Youtube, and Tiktok platforms regarding the use of BPA-free water gallons in Indonesia. From a total of 1200 data points, post-labeling, 772 data points were obtained, with 552 classified as positive and 220 as negative. The experimental results reveal that SVM achieves an accuracy of 96.15%, while Naïve Bayes achieves an accuracy of 84.55%. These findings indicate that SVM is effective in classifying sentiment with a high accuracy rate, providing valuable insights for manufacturers, government entities, and consumers regarding the use of BPA in water gallons in Indonesia. This study contributes to a better understanding of the role of social media in shaping public opinion and policies related to environmental and health issues.