Comparison of Machine Learning and Deep Learning Techniques for Identifying Hate Speech on Various Social Media Platforms Using Diverse Data Sets

Abstract

This study examine task is to evaluate a variety of machine learning techniques and methodologies for the purpose of detecting the appearance of hate speech on social media (SM).  This research was to investigate the essential components of hate speech classification using Machine Learning (ML) and Deep Learning (DL) techniques. Additionally, it explored the numerous obstacles that are experienced by different models. Generally speaking, the challenge of predicting hate speech is described as a task that involves categorizing text. It focused on five key areas such as feature extraction, dimensionality reduction, classifier development and selection, data exploration and collection, and model evaluation. Over time, the efficacy and efficiency of machine learning algorithms used to identify hate speech have significantly improved. There has been an influx of new performance measurements and datasets into the literature. A precise, thorough, and current state-of-the-art is required to educate researchers about new developments in automated hate speech identification. The findings of this study add up to three things. To begin, readers should be informed about the crucial procedures involved in hate speech identification utilizing machine learning algorithms. Second, the flaws and strengths of each technique are appraised to help researchers solve the algorithm Choice conundrum. Finally, significant research gaps and unsolved problems were discovered.