Identifikasi Kampanye Hitam Terhadap Calon Presiden Indonesia pada Pilpres Tahun 2024 di Media Sosial X Menggunakan Metode K-Means dan IndoBERT

Abstract

In this highly digital era, social media has become a widely used and favored platform for the public to interact, share information, and express their opinions. Among the various social media platforms available, one of the most popular in Indonesia is X. Ahead of the 2024 presidential election, the practice of black campaigns has increasingly emerged on various social media platforms, including X. Instead of focusing on the exchange of ideas promoted by each candidate, certain groups of presidential candidate supporters have begun spreading false information and hate speech. Black campaign attacks, consisting of false accusations, fake information, and hate speech, have been experienced by all presidential candidates contesting the 2024 presidential election. Therefore, a system is needed to effectively identify and detect tweets containing elements of black campaigns on social media X, enabling the public to filter campaign information related to each Indonesian presidential candidate without any indication of black campaigns. This will allow voters to make more informed and intelligent decisions when choosing their preferred candidate in the presidential election. The system developed in this study implements the K-Means algorithm and IndoBERT to identify and detect tweets containing elements of black campaigns targeting Indonesian presidential candidates in the 2024 presidential election. A total of 5,624 Indonesian-language tweet data were used in this study, achieving a model accuracy rate at 0.9796 or 97.96%.