Sentiment Analysis of Instagram Comments on the Ratification of the Criminal Procedure Code Bill using TF-IDF and Multinomial Naive Bayes

The ratification of the Draft Criminal Procedure Code (RUU KUHAP) has generated a broad and intense public response on social media. This study analyzes public perception of the ratification of the RUU KUHAP through Instagram user comments using Sentiment Analysis techniques based on TF-IDF representation and Multinomial Naive Bayes (MNB) classification. Comments were obtained using the Web Export method based on relevant post links, then preprocessed, converted into TF-IDF features, and classified. The evaluation showed an accuracy of around 66.67%, with the best performance in the negative class, and the positive class was not detected well due to unbalanced data distribution. The findings indicate the dominance of negative and neutral sentiments towards the RUU KUHAP. This study provides an empirical contribution to the understanding of public opinion on national legal issues.

Keywords: RUU KUHAP, sentiment analysis, Instagram, Multinomial Naive Bayes, TF-IDF