Deteksi Hate Speech pada Kolom Komentar TikTok dengan menggunakan SVM

Abstract

The TikTok application provides numerous features, including the comment section for users to interact with each other. Users can exchange their opinions openly through the comment section. However, as the interaction or exchange of opinions among users increases, the use of hate speech, consciously or unconsciously, remains prevalent. Hate speech refers to actions by an individual or group that can incite criminal acts, thereby harming others. This study aims to identify the use of hate speech in TikTok comment sections using the SVM algorithm and to compare two libraries used in the labeling process to observe the performance of the SVM algorithm model. The labeling process employs a lexicon-based approach. The dictionaries used in this study are the Inset lexicon and VaderSentiment. The SVM algorithm is used as the model to test the evaluation results. The results obtained using the Inset lexicon labeling show an accuracy of 82%, while the second labeling method using VaderSentiment yields an accuracy of 96.21%.