Klasifikasi Cyberbullying pada Media Sosial Tiktok Menggunakan Indobert Embedding dan Support Vector Machine

Abstract

TikTok is a social media platform that is free to use. This freedom has resulted in many users making various comments containing hate speech. This research aims to classify types of cyberbullying. This research uses the IndoBert Embedding and Support Vector Machine methods to create a machine learning model that can classify cyberbullying comments. This research uses Polynomial kernel with parameter C=10. The model is evaluated using F1 Score with cross validation technique to measure the model performance in cyberbullying classification. The results show that the Indobert Embedding and Support Vector Machine method can classify cyberbullying comments with an accuracy rate of 86% with cross validation evaluation and evaluation matrix.