Analisis Polarisasi Sentimen Kasus Bocornya Pusat Data Nasional Pada Platform Sosial Media

The National Data Center data leak case occurred in 2024. This triggered various public reactions, especially on social media and YouTube. Sentiment analysis was conducted to understand the various opinions about the data leak. This study aims to analyze the polarization of public sentiment related to this case to understand comments on data security issues. A total of 4,460 comments were collected from eight YouTube videos with different characters and classified into four categories: positive, negative, neutral, and others. Tests were conducted using IndoBERT and IndoBERTweet in three scenarios with customized parameters. The best results were obtained in the first scenario using IndoBERTweet with an accuracy of 0.87, while IndoBERT obtained a result of 0.86. Error analysis shows that misclassification often occurs due to words that are general and ambiguous, so the model has difficulty determining the true context of the comment.