PENDETEKSI UJARAN KEBENCIAN PADA PLATFORM MEDIA SOSIAL TWITTER MENGGUNAKAN SUPPORT VECTOR MACHINE

Abstract

Twitter is one of the world social media giants, which has enormous flow text-based comment in everysecond. There are many types of writing sentiments that users create to discuss something such as famousfigures, companies or politics. One of the types is hate speech. The analysis was carried out using theSupport Vector Machine method as a text analysis model with the help of TF-IDF to assess the weight ofeach word. The experiment was carried out with several types of kernels and resulted in varying degrees ofaccuracy. The types of kernels tested were linear, radial basis function, polynomial and sigmoid with a testdata distribution of 20%, 25% and 30%.