Hate Crimes Detection on Twitter using ML Techniques

With the sizeable adoption of social media structures which includes Twitter, the dissemination of hateful content material targeting people or agencies primarily based on race, gender, faith, or ethnicity has become increasingly commonplace. Manual moderation techniques are not scalable due to the sizable and swiftly developing extent of consumer-generated content material. This have a look at proposes a system studying-based framework to robotically stumble on and classify hate speech on Twitter. The pipeline entails comprehensive textual content preprocessing— normalization, tokenization, stopword elimination, and lemmatization— accompanied by using TF-IDF-primarily based characteristic extraction. Four type models—Logistic Regression, Support Vector Machine (SVM), Naive Bayes, and Random Forest—are evaluated the usage of a publicly to be had categorised Twitter dataset. Results suggest that SVM and Random Forest provide superior overall performance in terms of precision, don't forget, and basic accuracy. This paintings highlights the effectiveness of computerized methods in moderating dangerous on-line content and lays the foundation for destiny upgrades along with multilingual support and real-time detection.