Abusive Language Detection Using Auto-Machine Learning for Multiple Languages

Main Article Content

Abstract

News outlets and social media platforms struggle to find efficient and effective methods of moderating abusive online comments and posts. These comments can be extremely harmful to both users and businesses. Thus, abusive comment moderation is vital for the online comment section. Many companies do not have enough people to moderate comments manually; therefore, they need semi-automated comment moderation tools. Through this research, we explore the answer to the question of how Auto-Machine Learning can categorize and detect abusive English and German tweets. To begin, we will select data sets with multiple class labels, and we will develop different Auto-Machine Learning algorithms to see if a particular model fits the data sets well. To our knowledge, the use of Auto-Machine Learning in comment moderation remains untouched. Moreover, we will explore the approach of Auto-Machine Learning in comment moderation to assist the steps of processing the data sets, selecting the models, and efficiently detecting abusive language.

Article Details

Section
Research Reports

References

MIS Quarterly