Mapping Public Opinion on the DPR Salary Increase Issue via YouTube Comment Sentiment Analysis using IndoBERTa

Abstract

The issue of salary increases for members of the House of Representatives (DPR) in Indonesia has sparked widespread public debate because it is considered inconsistent with the socio-economic conditions of the community. The policy is perceived as sensitive, particularly regarding the principles of social justice, government accountability, and political legitimacy. In the digital era, social media such as YouTube has become an important space for the public to express opinions openly. This study aims to map public opinion regarding the DPR salary increase issue through sentiment analysis of YouTube comments using a Natural Language Processing (NLP) approach. The IndoBERTa model was used to classify public sentiment into positive, negative, and neutral categories, and n-gram analysis was used to capture dominant linguistic patterns. The results showed that negative sentiment dominated with 5,463 comments, far exceeding neutral (1,391 comments) and positive (812 comments). The n-gram analysis revealed that frequently appearing words and phrases related to "people," "DPR," "salary," as well as emotional expressions such as "disband the DPR" and "dancing on the people's suffering." These findings indicate that the DPR salary issue triggered a strong, often sarcastic public response and demonstrated a crisis of trust in the legislative institution. The lack of positive sentiment confirms that this policy has almost no public support.