Published Paper Details:

The proliferation of social media platforms has transformed global communication, enabling real-time sharing of information and connecting individuals across diverse geographies. However, this digital revolution has also facilitated the rise of cyberbullying, a pervasive issue that significantly impacts individuals' mental health and well-being. In the big data era, the sheer volume and velocity of data generated on social media present both challenges and opportunities for cyberbullying detection and prevention. This study explores the application of machine learning algorithms to predict and detect cyberbullying on social media platforms. Leveraging vast datasets from various social media sources, we employ advanced data preprocessing techniques, including natural language processing (NLP) for text normalization, sentiment analysis, and feature extraction. To address the challenge of imbalanced data, we utilize oversampling, undersampling, and synthetic data generation methods. We develop and compare multiple machine learning models, including supervised learning algorithms (Support Vector Machines, Logistic Regression, Random Forests) and deep learning architectures (Recurrent Neural Networks, Long Short-Term Memory networks, Convolutional Neural Networks, and Transformers). These models are evaluated using metrics such as accuracy, precision, recall, F1-score, and ROC-AUC to determine their effectiveness in detecting cyberbullying. Our findings demonstrate that machine learning algorithms, particularly deep learning models, exhibit high accuracy and precision in identifying cyberbullying patterns in text and multimedia content. We highlight the importance of continuous learning and model updating to adapt to evolving language and behavior on social media. This research underscores the potential of machine learning in enhancing the safety and security of online environments. By providing real-time monitoring and automated detection capabilities, these systems enable timely intervention and support for victims of cyberbullying. Additionally, the insights gained from analyzing social media interactions can inform the development of robust policies and educational programs aimed at preventing cyberbullying. The study also addresses ethical considerations, emphasizing the need for privacy-preserving techniques, bias mitigation, and compliance with legal standards. By advancing the field of cyberbullying detection through machine learning, we contribute to creating safer, more inclusive digital communities and promoting positive online interactions.