Analisis Komparatif Algoritma Naïve Bayes dan Support Vector Machine untuk Klasifikasi Sentimen Ulasan Pengguna RedNote di Google Play Store

Abstract

This study compares the performance of Naïve Bayes and Support Vector Machine (SVM) algorithms in classifying user sentiment toward the RedNote (Xiaohongshu) application based on reviews collected from the Google Play Store. A total of 2,364 reviews were collected through web scraping, comprising 1,373 positive reviews and 991 negative reviews. The research process involved text preprocessing, including case folding, non-alphabetic character removal, tokenization, stopword removal, and stemming. Text features were represented using Term Frequency–Inverse Document Frequency (TF-IDF), and the dataset was divided into training and testing sets using an 80:20 ratio. Model performance was evaluated using accuracy, precision, recall, and F1-score. The evaluation results showed that Naïve Bayes achieved an accuracy of 86%, outperforming SVM, which achieved 84% accuracy on the test set. Naïve Bayes also demonstrated a smaller gap between training and testing accuracy, whereas SVM exhibited a larger performance gap, suggesting a potential overfitting tendency. These findings indicate that Naïve Bayes is a more suitable model for the dataset and experimental setting examined in this study. This research provides a comparative evaluation of two machine learning algorithms for classifying RedNote user reviews and highlights the importance of assessing model performance beyond training accuracy.