Detection of hate speech using machine learning algorithms

The rise of social media and sharing knowledge is good for humanity. But it’s come with other problems, like the spread and amplification of hate speech. Hate speech remains a nagging problem on social media sites. As users generate online content so rapidly, it has rendered manual moderation cumbersome and hard to process for user comments on hate speech, since doing would dilute the blow of nourishing the latter. Existing research has successfully built models to identify the occurrences of hate speech in user comments and tweets on the social media network Twitter. These models provide good results though when applied to the real world, perform poorly on tweets that contain lexical peculiarities such as smart language and novel forms of spelling or are written with terms seldom regarded as an indicator of obscene content. In order to tackle these challenges, it is necessary to develop a solution that will have the ability to identify hatred in comments made by social media users. This study will explore different machine learning algorithms and methods for identification of hate speech in social media. This research investigated the essentials for hate speech categorization. The five basic baseline components that were analyzed: data collection and exploration, feature extraction, dimensionality reduction, classifier selection and training as well as model evaluation. Hate Speech Detection is a natural language processing task that has evolved over time through machine learning methods. With all these advances taking place automatically detecting hate speech is a hot research area demanding a thorough and timely state-of-the-art to keep researchers updated on the progress made to date in the automatic detection of hate speech.