USING MACHINE LEARNING TO INSTANTLY DETECT HATE SPEECH ON SOCIAL MEDIA

The prevalence of hate speech on social media has necessitated the development of robust methods for its detection to maintain online civility and safety. In this research paper, the methods and materials section explored the use of multiple machine learning algorithms to detect hate speech using different feature sets such as statistical data, term frequency-inverse document frequency (TFIDF), and linguistic analysis and word count (LIWC). The study compared the performance of six well-known machine learning models — Random Forest, Decision Tree, K-Nearest Neighbors (KNN), Naive Bayes, Logistic Regression, and Support Vector Machines — on criteria such as precision, recall, F-measure, and area under the curve (AUC-ROC) values. Scientific novelty of this research work is enhanced hate speech detection using hybrid feature engineering and ethical moderation framework. Practical significance of the work is improves hate speech detection accuracy for safer social media while ensuring ethical content moderation. The results showed that models using a mixture of complex linguistic and statistical attributes significantly outperformed those relying on a basic feature set. This highlighted the critical importance of fine-grained feature engineering in the detection process. The study also considered the ethical implications of automated hate speech detection, highlighting the need for balanced strategies that maintain the effectiveness of content moderation while protecting free speech. This study contributed to the field by identifying both the strengths and weaknesses of existing methodologies. It suggested avenues for future improvement, including the integration of advanced natural language processing techniques and continually raising ethical standards when implementing such models.