Analisis Sentimen Pengguna X Terhadap Kebocoran Data Pribadi Menggunakan Algoritma Naïve Bayes Classifier

Abstract

Personal data breaches in Indonesia have become increasingly prevalent, posing significant risks to both individuals and businesses. With the rapid advancement of information technology, this issue has sparked intense discussions on social media, particularly on the X (Twitter) platform. This study aims to analyze user sentiment on X regarding personal data breaches by classifying opinions into positive, negative, and neutral sentiment categories. Additionally, this research evaluates the performance of the classification model using a confusion matrix to measure accuracy, precision, recall, and f1-score. The method used in this study is the Naïve Bayes Classifier, with 70% training data (433 data points) and 30% testing data (186 data points). The data was obtained through web crawling and preprocessed before performing sentiment classification. The results indicate that negative sentiment dominates with 43.54%, followed by positive sentiment at 28.50%, and neutral sentiment at 27.96%. Model evaluation achieved an accuracy of 98.92%, with negative precision 100%, neutral precision 100%, and positive precision 96.30%. Meanwhile, recall for both positive and negative sentiment reached 100%, while recall for neutral sentiment was 96.23%. The f1-score for negative, neutral, and positive sentiment was 1.0, 0.988, and 0.981, respectively. These findings demonstrate that the Naïve Bayes Classifier performs exceptionally well in classifying sentiment related to personal data breaches. The dominance of negative sentiment in the classification results reflects high public concern over this issue, highlighting the urgent need for enhanced data security measures and privacy protection in Indonesia.