Aplikasi Pendeteksi Unsur Hinaan dalam Komentar di Media Sosial Berbahasa Indonesia

Yohanes Adam Sastrodikoro Wiryoadikusmo (Program Studi Teknik Informatika)
Henry Novianus Palit (Program Studi Teknik Informatika)
Justinus Andjarwirawan (Program Studi Teknik Informatika)

Jurnal Infra
Vol 6, No 1 (2018)

Communication in social media has became a lifestyle of Indonesian people. Most of the kind of communication between users are commentary text. Often there are comments that contain elements of insults directed at others. Detection of insult elements in comment text in social media is needed.The first step in the method is to normalize the text so that the comment text is more organized and standardized. The replacement of the insult verbal variation into the original insult word is also done. Then do the stemming process to take the basic word of each word. By performing feature extraction of the text, obtained every feature of the text is then predicted using the SVM method.The application used in this paper successfully detected insulting elements in the text of comments from social media using the three methods of SVM classification, Logistic Regression, and Naive Bayes.