Abstract

The increasing use of social media and digital communication tools has brought many advantages, yet it has also enabled the concerning issue of cyberbullying which involves the use of digital platforms to harass, threaten, dishonour or demean individuals. Techniques to reduce this problem to a certain extent have already been introduced in online systems. However, they possess limitations of usage and lack of explainability. This research presents the development and evaluation of an Explainable Artificial Intelligence (XAI) based cyberbullying classification system intended to detect and interpret harmful content on social media platforms. The system leverages textual and symbolic features from user-generated contents, extracted from social media platform, X (Twitter), to classify posts as bullying or non-bullying. The methodology adopted in this work include data preprocessing (tokenization, stemming, lemmatization, and TF-IDF vectorization), followed by model training using Logistic Regression (LR), Decision Tree (DT), and Multinomial Naive Bayes (MNB) classifiers. To address the class imbalance in the dataset, Synthetic Minority Over-sampling Technique (SMOTE) was employed, which resulted in improved model fairness and performance. Among the models that were tested, Logistic Regression achieved the highest generalization accuracy of 93.12%. In other to enhance transparency and trust, SHAP (SHapley Additive exPlanations) was integrated, which offered interpretable insights into model predictions and highlights key linguistic features influencing classification results. The system was deployed via a web interface which enables real-time content moderation. While the results demonstrate high accuracy and interpretability, the study also reflects ethical considerations such as contextual misclassification, societal bias, and data privacy.

References

• Martin, F., Zhuang, M., & Schaefer, D. (2024). Systematic review of research on artificial intelligence in K-12 education (2017–2022). Computers and Education: Artificial Intelligence, 6, 100195.
• Gupta, A. K., Kumar, A., & Kumar, B. (2025). Advancing sustainable food packaging: Integrating machine learning, deep learning, and artificial intelligence. Trends in Food Science & Technology, 163, 105148.
• Sharma, A., Lysenko, A., Jia, S., Boroevich, K. A., & Tsunoda, T. (2024). Advances in AI and machine learning for predictive medicine. Journal of Human Genetics, 69(10), 487-497.
• Saifullah, K., Khan, M., Jamal, S. and Sarker, I (2024). Cyberbullying Text Identification based on Deep Learning and Transformer-based Language Models. EAI Endorsed Transactions on Industrial Networks and Intelligent Systems, 11, 1–12.
• Prama, T. T., Amrin, J. F., Anwar, M. M., & Sarker, I. H. (2025). AI Enabled User-Specific Cyberbullying Severity Detection with Explainability. arXiv preprint arXiv:2503.10650.
• Aronna, S. R., Zumma, T., Prodhan, R., Zohora, F., Sakib, N. and Tahmiduzzaman, K. (2023). A Study of Cyber Bullying Classification Using Social Media and Textual Analysis Based on Machine Learning Approaches. In Proceeding of 14th ICCCNT IEEE Conference, IIT, Delhi, India.
• Garzia-Mendez, S. and Arriba-Perez, F. (2025). Promoting Security and Trust on Social Networks: Explainable Cyberbullying Detection using Large Language Models in a Stream-based Machine Learning Framework. arXiv:2505.03746v1 [cs.SI].
• Armas, D. G. A., Toapanta, S. M. T., Díaz, E. Z. G., Guerrero, J. L. J., Arellano, M. R. M., & Hifóng, M. M. B. (2025). Influence of Social Media and Artificial Intelligence on Cyberbullying for Decision-Making with Legal or Judicial Foundations in Ecuador. Journal of Internet Services in Information Security 15(1), 32-50.
• Perera, A., & Fernando, P. (2024). Cyberbullying detection system on social media using supervised machine learning. Procedia Computer Science, 239, 506-516.
• Unnava, S., & Parasana, S. R. (2024). A study of cyberbullying detection and classification techniques: A machine learning approach. Engineering, Technology & Applied Science Research, 14(4), 15607-15613.
• Ambareen, K., & Meenakshi Sundaram, S. (2023). A Survey of Cyberbullying Detection and Performance: Its Impact in Social Media using Artificial Intelligence. SN Computer Science, 4(6), 859.
• Xu, Q., Feng, Z., Gong, C., Wu, X., Zhao, H., Ye, Z., ... & Wei, C. (2024). Applications of explainable AI in natural language processing. Global Academic Frontiers, 2(3), 51-64.
• Uddin, M. K. S. (2024). A review of utilizing natural language processing and AI for advanced data visualization in real-time analytics. Global Mainstream Journal, 1(4), 10-62304.
• Sarella, P. N. K., & Mangam, V. T. (2024). AI-driven natural language processing in healthcare: transforming patient-provider communication. Indian Journal of Pharmacy Practice, 17(1), 21-26.
• Krugmann, J. O., & Hartmann, J. (2024). Sentiment analysis in the age of generative AI. Customer Needs and Solutions, 11(1), 3.
• Fashakh, A. M., Çevik, M., Aydoğan, Ş. K., & Ibrahim, A. A. (2025). Detection cyberbullying using AI and sentiment analysis to examine psychological impacts on vulnerable groups. Egyptian Informatics Journal, 32, 100856.
• Iwendi, C., Srivastava, G., Khan, S., & Maddikunta, P. K. R. (2023). Cyberbullying detection solutions based on deep learning architectures. Multimedia Systems, 29(3), 1839-1852.
• Abood, M. M., & Al-Bayati, M. A. (2024). Explainable Multimodal Deep Learning Model for Cyberbullying Detection (EMDL-CBD). Journal Port Science Research, 7(3), 268 – 280
• Ajayi, E., Kachweka, M., Deku, M., & Aiken, E. (2025). A Machine Learning Approach for Detection of Mental Health Conditions and Cyberbullying from Social Media. arXiv preprint arXiv:2511.20001.
• Wulandari, W., Makmur, H., Surianto, D. F., Risal, A. A. N., Budiarti, N. A. E., Zain, S. G., & Wahid, A. (2025). Semantic Feature Engineering with LSA-SVM for Cyberbullying Comment Classification on instagram, Informatica, 49(15), 165 – 178.
• Zhao, Z. (2025).0 Let Network Decide What to Learn: Symbolic music understanding model based on large-scale adversarial pre-training. In Proceedings of the 2025 International Conference on Multimedia Retrieval (ICMR ’25), June 30-July 3, 2025, Chicago, IL, USA. ACM, New York, NY, USA,
• Bennetot, A., Donadello, I., El Qadi El Haouari, A., Dragoni, M., Frossard, T., Wagner, B., ... & Diaz-Rodriguez, N. (2024). A practical tutorial on explainable AI techniques. ACM Computing Surveys, 57(2), 1-44.
• Tilala, M. H., Chenchala, P. K., Choppadandi, A., Kaur, J., Naguri, S., Saoji, R., ... & Tilala, M. (2024). Ethical considerations in the use of artificial intelligence and machine learning in health care: a comprehensive review. Cureus, 16(6), e62443. doi: 10.7759/cureus.62443
• Petch, J., Di, S., & Nelson, W. (2022). Opening the black box: the promise and limitations of explainable machine learning in cardiology. Canadian Journal of Cardiology, 38(2), 204-213.
• Ige, T., & Adewale, S. (2022). AI powered anti-cyber bullying system using machine learning algorithm of multinomial naïve Bayes and optimized linear support vector machine. arXiv preprint arXiv:2207.11897.
• Tasin, I., Nabil, T. U., Islam, S., & Khan, R. (2023). Diabetes prediction using machine learning and explainable AI techniques. Healthcare technology letters, 10(1-2), 1-10.
• Almufareh, M. F., Jhanjhi, N. Z., Humayun, M., Alwakid, G. N., Javed, D., & Almuayqil, S. N. (2025). +Integrating sentiment analysis with machine learning for cyberbullying detection on social media. IEEE Access, 13, 78348-78359.
• Sanchez, R., Hernández, P., & Fernandez, E. (2023). Integrating RNNs with rule-based explanations for temporal pattern recognition in cyberbullying. Applied Intelligence, 53(4), 3562-3578.
• Shah, V., Sinha, A., Navalkar, N., Gupta, S., Gonsalves, P., & Malik, A. (2023). ML and Natural Language Processing: Cyberbullying Detection System for Safer and Culturally Adaptive Digital Communities. Journal of Smart Internet of Things, 2, 193-205.
• Krak, I., Sobko, O., Molchanova, M., Tymofiiev, I., Mazurets, O., & Barmak, O. (2024). Method for Neural Network Cyberbullying Detection in Text Content with Visual Analytic. In CS&SE@ SW, 298-309.