TINJAUAN PUSTAKA SISTEMATIS: KLASIFIKASI UJARAN KEBENCIAN PADA SOSIAL MEDIA DENGAN ALGORITMA NEURAL NETWORK

Abstract

Hate speech is illegal because it incites violence and anarchic attitudes towards other people or groups. Hate speech includes words, behavior, and actions. The importance of social media ethics is emphasized because the internet is seen as an important component of modern society. However, more and more parties are abusing the internet to spread information about racial, religious and ethnic hatred. This is something that needs attention because of the proliferation of hate speech on the internet. To find a pattern, one method that can be used is Machine Learning (ML). Text is one of the data types that ML (known as text analytics) can apply to. Previous research has used the Support Vector Machine (SVM) technique to find hate speech on social media that has more than one label (multilabel). The Neural Network Algorithm is used in this study to identify hate speech on social media content that has multiple labels (multilabel). The multi-label hate speech dataset in Indonesian text is used in this study, and the results show that the Neural Network model performs best, with an accuracy rate of 98.76 percent.