Analisis Perbandingan Algoritma Naïve Bayes dan Support Vector Machine dalam Klasifikasi Opini dan Fakta pada Berita Banjir Sumatera

The rapid growth of online media has accelerated the dissemination of information related to flood disasters, creating the need for an automatic method to classify news into factual and opinion categories. This study aims to compare the performance of the Naïve Bayes and Support Vector Machine (SVM) algorithms in classifying factual and opinion-based news on flood events in Sumatra using Term Frequency–Inverse Document Frequency (TF-IDF) weighting. The research employed several text preprocessing stages, including cleaning, case folding, tokenization, stopword removal, and stemming, followed by TF-IDF weighting, classification, and model evaluation using accuracy, precision, recall, and F1-score. The experimental results showed that the Naïve Bayes algorithm achieved an accuracy of 92.34%, outperforming the Support Vector Machine algorithm, which achieved an accuracy of 91.88%. In addition, Naïve Bayes obtained higher precision, recall, and F1-score values for the opinion class. These findings indicate that Naïve Bayes is more effective than Support Vector Machine in classifying factual and opinion-based news related to flood events in Sumatra.