The unprecedented growth of social media platforms has transformed online communication while simultaneously amplifying the spread of hate speech and abusive content. Facebook, as the world’s largest social networking platform, presents unique challenges for automated content moderation due to its diverse user base, longer text formats, conversational context, and multilingual usage. Machine learning (ML) and natural language processing (NLP) techniques have emerged as scalable solutions for detecting hate speech; however, their effectiveness varies significantly depending on the chosen algorithms and data characteristics. This paper presents a comprehensive and original analysis of widely used machine learning and deep learning algorithms for hate speech detection on Facebook. Traditional classifiers such as Naïve Bayes, Logistic Regression, Support Vector Machines, Random Forests, and Gradient Boosting are examined alongside deep learning models including Long Short-Term Memory (LSTM) networks and transformer-based architectures such as BERT. Mathematical formulations, algorithmic workflows, and comparative performance analyses are provided to highlight the strengths and limitations of each approach. The paper establishes a strong methodological foundation for subsequent bias-aware and sentiment-integrated hate speech detection frameworks.
Singhi et al. (Mon,) studied this question.