ANALISIS SENTIMEN TERHADAP PIDATO PRESIDEN 2026 PADA APLIKASI TIKTOK DAN X MENGGUNAKAN METODE NAIVE BAYES CLASSIFIER

Abstract

The rise of social media platforms like TikTok and X has made them primary channels for the public to express opinions on various public issues, including the 2026 Presidential speech. The massive volume of comments generated necessitates sentiment analysis to automatically identify public opinion trends. This study aims to analyze public sentiment regarding the 2026 Presidential speech on TikTok and X using the Naive Bayes Classifier method. A quantitative approach based on text mining was employed, utilizing a secondary dataset of 2,000 comments obtained from Kaggle. The data underwent pre-processing stages—including cleaning, case folding, tokenizing, stopword removal, and stemming—followed by feature extraction using Term Frequency–Inverse Document Frequency (TF-IDF). Classification was performed using the Multinomial Naive Bayes algorithm and evaluated via a confusion matrix based on accuracy, precision, recall, and F1-score. The results indicate a sentiment distribution of 47.00% positive, 29.90% neutral, and 23.10% negative. The developed model achieved an accuracy of 82.25%, a weighted precision of 84.22%, a weighted recall of 82.25%, and a weighted F1-score of 82.06%, demonstrating good classification performance on Indonesian text data. However, the model struggled to distinguish between neutral and negative sentiments due to linguistic context similarities. The study concludes that the combination of pre-processing, TF-IDF weighting, and the Naive Bayes Classifier algorithm constitutes an effective approach for social media sentiment analysis; nevertheless, validation using original scraped data and comparisons with other algorithms are still required to enhance the model's accuracy and generalization capabilities