Automated Hate Speech Detection on Vietnamese Social Networks

EasyChair Preprint no. 1745

Abstract

Nowadays, the internet plays an important role in our everyday life. It provides us useful information, knowledge, news, and a free space to share and exchange our personal opinions with other people all over the world through some platforms such as the social network. While there are various advantages of social network, its freedom sometimes bring to us a lot of trouble. Some people use the social network for some immoral aims such as harass, racist, and offend others which must be detected and removed immediately. With the rapid development of the social network, number of content uploaded on it is dramatically enormous and becoming larger which can not control effectively by human. We proposed a novel method for solving this problem by a multi-class classification model to classify content into 3 labels: HATE, OFFENSIVE, and CLEAN. With the Vietnamese dataset of the competition VLSP-SHARED Task, our experimental results have the first position on the contest table.

Keyphrases: Automated Hate Speech Detection, hate speech detection, Natural Language Processing, text mining, Vietnamese Hate Speech Detection

@Booklet{EasyChair:1745,
  author = {Quang Pham Huu and Son Nguyen Trung and Hoang Anh Pham},
  title = {Automated Hate Speech Detection on Vietnamese Social Networks},
  howpublished = {EasyChair Preprint no. 1745},

  year = {EasyChair, 2019}}