Research Article
A Unified Adaptive Cyber Threat Intelligence Model for Real-Time IoT Security Using Machine Learning and GAN-Based Augmentation
Issue:
Volume 13, Issue 3, September 2025
Pages:
52-61
Received:
13 August 2025
Accepted:
25 August 2025
Published:
13 September 2025
Abstract: The rapid rise of Internet of Things (IoT) devices has made cybersecurity much more dangerous and vulnerable, emphasizing the critical necessity for adaptive intrusion detection systems (IDS) to safeguard IoT networks. This study presents a Cyber Threat Intelligence (CTI) model that works in real time and adapts to IoT contexts. The suggested model uses density-based clustering (DBSCAN), deep learning (CNN-LSTM), and reinforcement learning (LDQN) to find, sort, and respond to threats that change over time. A generative model (GAN) is added to make detection better by adding fake data. The model works in three main steps: detection, mitigation and response, and ongoing improvement which is adaptively. During the detecting phase, DBSCAN identifies anomalies by grouping network IoT traffic and separating outliers. A hybrid CNN-LSTM architecture processes anomalies by finding patterns of threats over time, while a Random Forest algorithm classifies typical traffic. During the mitigation and response phase, a Lightweight Deep Q-Network (LDQN) dynamically assigns the actions BLOCK, DROP, INVESTIGATE, or ALLOW based on how serious each threat is. A Generative Adversarial Network (GAN) produces fake data to fix class imbalance and make it easier to find classes that aren't well represented. After being improved, the unified model was able to find IoT intrusions with an accuracy of 92.86%, a precision of 95.16%, and a recall of 95.93%. The system learns about new attack patterns in real time and responds to threats automatically, making it useful for protecting big and changing IoT deployments. This research links classic IDS solutions with cutting-edge AI-driven threat intelligence systems to create an approach for IoT cybersecurity that can grow, is resilient, and improves itself.
Abstract: The rapid rise of Internet of Things (IoT) devices has made cybersecurity much more dangerous and vulnerable, emphasizing the critical necessity for adaptive intrusion detection systems (IDS) to safeguard IoT networks. This study presents a Cyber Threat Intelligence (CTI) model that works in real time and adapts to IoT contexts. The suggested model...
Show More
Research Article
A Privacy-Preserving Data Governance in Cross-Border Telemedicine Using Federated Learning and Differential Privacy in Kenya
Issue:
Volume 13, Issue 3, September 2025
Pages:
62-76
Received:
19 August 2025
Accepted:
1 September 2025
Published:
19 September 2025
Abstract: This study presents a privacy-preserving learning model designed for cross-border telemedicine in East Africa that keeps raw patient records in country while hospitals collaborate on model quality. The core of this approach is to keep sensitive patient records localized within each country, with hospitals training models locally and only sharing model updates. Using synthetic EHRs split across seven hospitals in Kenya, Tanzania, and Uganda, we compare centralized training, standard federated learning, and federated learning with differential privacy. Federated learning improves utility while maintaining data localization, with accuracy rising by about 0.0665, recall for the positive class improving by about 0.1193, and F1 increasing by about 0.0657 relative to centralized training. Adding differential privacy made the system more resilient to attacks. The success rate of model-inversion attacks dropped from 0.696 in the centralized training scenario to 0.686 with standard FL and further to 0.638 with FL + DP. This represents an absolute reduction of 0.058, or about 8.4 percent, in attack success. Membership-inference leakage has an AUC of around 0.50. The trade-off is tunable utility at a chosen privacy budget, for example accuracy near 0.530 at ε = 0.30. The originality is practical, we pair federated learning with an attack simulator and an ε register that turns privacy into an auditable setting hospitals can manage during cross-border care.
Abstract: This study presents a privacy-preserving learning model designed for cross-border telemedicine in East Africa that keeps raw patient records in country while hospitals collaborate on model quality. The core of this approach is to keep sensitive patient records localized within each country, with hospitals training models locally and only sharing mo...
Show More