Adaptive Privacy-Preserving Federated Learning Framework For Heterogeneous IOT Environments

The dataset used in this study for the proposed “Adaptive Privacy-Preserving Federated Learning Framework for Heterogeneous IoT Environments” consists of heterogeneous IoT network and healthcare-related traffic data collected from distributed edge devices and publicly available benchmark repositories. The dataset includes multiple attributes related to device communication patterns, sensor-generated data, encrypted traffic characteristics, network behavior, and security-related events to support federated learning-based privacy-preserving analysis. The data were preprocessed through normalization, feature extraction, missing value handling, and noise reduction techniques to improve model performance and interoperability across heterogeneous IoT environments. The dataset was utilized to evaluate the framework’s efficiency in secure decentralized learning, privacy preservation, adaptive model aggregation, and intelligent threat detection under real-time distributed IoT scenarios.