Description:

In the digital age, ensuring user safety on social media platforms like Twitter is crucial due to the prevalence of cyberbullying, which affects approximately 37% of adolescents and poses significant mental health risks. To combat this, we aim to develop an intelligent system using NLP techniques to identify and flag cyberbullying instances. Our approach involves training machine learning models—Logistic Regression, Bernoulli Naive Bayes (BernoulliNB), and Linear Support Vector Classifier (LinearSVC)—on a dataset containing 1.6 million rows of Twitter conversations, utilizing TF-IDF vectorization to analyze and differentiate harmful behavior from harmless discourse.