Sentiment Analysis of Hate Speech Against Presidential Candidates of the Republic of Indonesia in the 2024 Election Using BERT

Abstract

The issue of hate speech on social media has become a matter of growing concern, particularly in the context of political discourse, as evidenced by the 2024 elections in Indonesia. Online platforms such as YouTube represent a primary medium for political discourse, frequently accompanied by negative or hateful commentary directed towards presidential candidates. The objective of this study is to analyze the sentiment of YouTube comments related to Indonesian presidential candidates in the 2024 General Election using the BERT algorithm. The data was obtained through scraping using the YouTube API and subsequently categorized into three distinct categories of hate speech: The categories of hate speech are as follows: OFP (offensive personal), OFG (offensive group), and OFO (offensive others). The CRISP-DM method was employed in this research, which included the following stages: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. The results demonstrate that the BERT algorithm is capable of classifying comments with a satisfactory level of accuracy. This algorithm can be utilized to develop predictive applications that assist in identifying and managing hate speech on social media.