Pemanfaatan Metode Klasifikasi Naïve Bayes Untuk Pendeteksi Berita Hoax Pada Artikel Berbahasa Indonesia

Abstract

– Hoax news has been widely spread on the internet. Ease of creating and sharing is one factor. Hoax news is a threat and concentration of many parties, problems arise in identifying or classifying them because there is no identifiable pattern, and the writing style is free and not rigid. The lack of accuracy of the existing hoax detection system found the methods and attributes used to classify hoax news with high accuracy. On this basis, this research was carried out, as in most classifications of hoax news used as reference in this study, preprocessing was carried out (case folding, tokenization, stemming, and stopword removal), feature extraction and adding attributes other than preprocessing articles. done like the website where the article is posted. publication and site status. The results of this study obtained an accuracy of 72% which turned out to be a decrease of 6.6% compared to previous research, namely 78.6% because one site only publishes one hoax article and allows the site's domain to expire thereby reducing the weight of the classification value.