A Comprehensive Study on the Effectiveness of Machine Learning Models to Mitigate the Impact of Cyberbullying
Publication Date
3-9-2026
Document Type
Conference Proceeding
Publication Title
2026 International Conference on Computing Networking and Communications Icnc 2026
DOI
10.1109/ICNC68183.2026.11416855
First Page
367
Last Page
371
Abstract
Cyberbullying affects 20-40% of digitally connected adolescents, yet existing automated detection systems achieve highly variable accuracy and lack ethical decision-making mechanisms. This study presents a unified AI framework combining Machine Learning (ML), Deep Learning (DL), and hybrid models to address cyberbullying detection across four major platforms: Wikipedia Talk, Twitter, Facebook, and YouTube. The framework introduces a novel confidence-based moderation layer that converts prediction certainty into three actionable tiers: automatic blocking (high confidence), user warnings (moderate confidence), and content publishing (low confidence). By addressing persistent challenges in platform adaptability, model interpretability, and ethical decision-making, the proposed framework provides a scalable and robust solution for real-time cyberbullying detection that strikes a balance between high technical performance and responsible, real-world deployment.
Keywords
Convolutional Neural Network, Cyberbullying
Department
Computer Science; Mechanical Engineering; Aviation and Technology
Recommended Citation
Shweta Singh, Fred Barez, and Riti Gour. "A Comprehensive Study on the Effectiveness of Machine Learning Models to Mitigate the Impact of Cyberbullying" 2026 International Conference on Computing Networking and Communications Icnc 2026 (2026): 367-371. https://doi.org/10.1109/ICNC68183.2026.11416855
