Hoax Web Detection For News in Bahasa Using Support Vector Machine

Abstract

This research creates a web-based user-friendly system that aims to detect hoax and non-hoax news of Indonesian language news links. The input data is in the form of links and archive sites from the Forum Anti Fitnah Hasut dan Hoax (FAFHH), using 100 news for training data and 20 news for test data that is processed by crawling and then processed in the pre-processing phase, namely tokenizing, stop word and stemming. Next is the Term Frequency-Inverse Document Frequency (TF-IDF) stage to provide weighting data which will be input data at the classification stage using the Support Vector Machine (SVM) Algorithm with a linear kernel to detect hoax and non-hoax news. The experimental results show that the system can classify well with an accuracy of 85%.