A Classification Model Based on Machine Learning for Detecting Racist Comments on Social Media Platforms

Abstract

Racial conflicts have become even more prevalent than before. As a result, social media companies are continuously being slammed for their inadequate response to the problem caused by racial discrimination. For example, the year 2020 witnessed a worldwide movement calling for racial equality and justice. The movement began after an African American male was suffocated and murdered by an NYPD police officer. Since then, there has been significant research efforts focusing on social media and the role it has played in amplifying racism. Similarly, the government of the United Kingdom have threatened to make social media companies legally accountable for the racist content on their platform after the witnessed increase of racist abuse on footballers in 2021. English football clubs have also threatened a boycott of social media in a bid to eradicate online hate. To solve this problem, we will track down past events and social media trends which are likely to have triggered racist reactions and retrieve annotated comments from public social media sites like Facebook, Instagram, Twitter, YouTube and TikTok. We will create an unbiased dataset of racist comments across social media platforms. We will be building a classification model using machine learning to detect racist comments on social media platforms. We propose a machine learning model for the automatic detection of racist comment across social media platforms. The results we obtained from our research shows that the support vector machine-trained model performs the best with an accuracy of 88.19%. The models proposed in this research outperformed most of the pre-existing models for the same task. Keywords: Race, Racism, Cyberbully, Hate Speech, Support Vector Machine, Confusion Matrix Journal Reference Format: Allenotor, D. & Oyemade, D. A. (2021): A Classification Model Based on Machine Learning for Detecting Racist Comments on Social Media Platforms Journal of Behavioural Informatics, Digital Humanities and Development Research. Vol. 7.No. 1, Pp 121-136. ICT University USA Endowed Research Series Publication in collaboration with SMART-Africa. Available online at https://www.isteams.net.behavioraljournal. Article DOI No - dx.doi.org/10.22624/AIMS/BHI/V7N1P9