Deep Learning-Based Cybersecurity Threat Detection and Vulnerability Prediction Model for Ethiopian Higher Education Institutions: A Comparative Case Study of MAU, AASTU, and BDU

The rapid digitalization of Ethiopian higher education institutions has exposed critical cybersecurity vulnerabilities that threaten data integrity and institutional operations. Over a 12-month period, this research addresses these gaps by developing a deep learning based threat detection and vulnerability prediction model through a comparative case study of Mekdela Amba University, Addis Ababa Science and Technology University, and Bahir Dar Institute of Technology. Employing a multi-modal approach that integrates 3 months of network log analysis, behavioral phishing simulations, and security awareness surveys, the project will train a hybrid CNN-LSTM-Transformer model capable of threat detection while incorporating institutional security policies as configurable parameters. The initiative aims to deliver an accurate, policy-aware cybersecurity prediction system, a comparative threat landscape analysis, and an adaptable national framework, ultimately strengthening cybersecurity postures across Ethiopia’s higher education sector within the project timeframe and providing a scalable, cost-effective model for future AI-driven security research in developing educational contexts.