Public Sentiment Analysis on Bullying Cases of Binus Serpong Students Using the Naive Bayes Method

Abstract

Sentiment analysis has evolved into a vital tool for comprehending public perceptions of social issues, including bullying cases in educational settings. This research focuses on sentiment analysis of public perceptions regarding bullying cases involving students at High School Binus Serpong. The dataset used consists of 2,259 comments collected from YouTube, categorized into three sentiment types : positive, negative, and neutral. The method applied in this study is the Complement Naive Bayes algorithm and multinomial Naive Bayes integrated with TF-IDF within a classification pipeline. The analysis results indicate that the classification model achieved an accuracy of 80.47%. Specifically, the evaluation results show that negative sentiment contributes 72.2%, positive sentiment 20.7%, and neutral sentiment 0.7%. Furthermore, the evaluation results indicate that negative sentiment has a precision of 0.81, recall of 0.93, and f1-score of 0.87, positive sentiment has a precision of 0.67, recall of 0.40, and f1-score of 0.50; while neutral sentiment has a precision of 0.70, recall of 0.43, and f1-score of 0.53.