Analisis Sentimen Pengguna Media Sosial Tiktok Terhadap Kinerja Wakil Presiden Gibran Rakabuming Raka Menggunakan Metode Naive Bayes

Abstract

TikTok social media generates a large volume of unstructured textual comments that can be utilized as a source for sentiment analysis using machine learning methods. This study focuses on the sentiment analysis of TikTok users' comments regarding the performance of Vice President Gibran Rakabuming Raka by applying the Naive Bayes algorithm as the classification method. Data were collected through a scraping technique from comments on TikTok videos published during the 2025–2026 period. A total of 7,249 comments were successfully collected, and after the selection and preprocessing stages, 4,800 comments were retained as the research dataset, consisting of 80% training data and 20% testing data. The data preprocessing stages included case folding, cleaning, tokenizing, stopword removal, stemming, and feature extraction using the Term Frequency–Inverse Document Frequency (TF-IDF) weighting method, followed by sentiment classification using the Naive Bayes algorithm. The model performance was evaluated using a confusion matrix by measuring accuracy, precision, recall, and F1-score. The experimental results showed that 2,546 comments (53.04%) were classified as negative sentiment, while 2,254 comments (46.96%) were classified as positive sentiment. The model achieved an accuracy of 87%, with a precision of 85% for the negative class and 90% for the positive class, a recall of 92% for the negative class and 81% for the positive class, and F1-scores of 88% and 85% for the negative and positive classes, respectively. These findings indicate that the Naive Bayes algorithm is effective in classifying the sentiment of TikTok users' comments with a high level of accuracy