Multilabel Classification for Toxic Comments in Indonesian

Authors

  • Reinert Yosua Rumagit Bina Nusantara University

DOI:

https://doi.org/10.21512/emacsjournal.v2i1.6256

Abstract

The more rapid development of the internet world, users can make comments on a variety of content on social networks, such as social media, blogs and others. Free users make comments triggering negative comments, making insults and incitement. By classifying user comments it is hoped that the system can be smarter to be able to distinguish threat, insult and incitement comments. The technique for classifying user comments uses deep learning, consisting of 6 classes. The results of experiments that have been conducted show that deep learning models produce an accuracy rate above 98%.
Dimensions

Plum Analytics

Author Biography

Reinert Yosua Rumagit, Bina Nusantara University

Computer Science Department, School of Computer Science, Bina Nusantara University

Jakarta, Indonesia 11480

Downloads

Published

2020-01-28

Issue

Section

Articles
Abstract 314  .
PDF downloaded 507  .