Heart Disease Detection and Prognosis Using IoT-Based ECG Sensor Data with Hybrid Deep Learning Architecture and Optimal Resource Allocation

Heart disease remains a major global cause of mortality, underscoring the need for advancements in early detection and prognosis to enhance patient recovery. This study proposes an innovative framework integrating deep learning (DL) models and optimal resource allocation strategies to improve heart disease detection and prognosis. The framework begins with rigorous preprocessing of the Internet of Things (IoT) captured Electrocardiogram (ECG) data, employing min–max normalization, and advanced median filtering techniques for noise reduction and baseline wander correction. Statistical features are extracted from the preprocessed data, while features such Improved Empirical Mode Decomposition (EMD), RR interval, R peak, and PR interval are derived from ECG signals. These features are then augmented using a min–max technique to enhance dataset diversity and model robustness. Furthermore, the study introduces a hybrid model combining a Deep Residual Network (DRN) and an Improved Bidirectional Gated Recurrent Unit for severity classification heart disease detection, leveraging augmented features. Optimal resource allocation is facilitated by the Improved Walrus Optimization Algorithm (WaOA), optimizing ventilator, Intensive Care Unit (ICU) bed, medical staff, and medication allocation based on predicted disease severity. Evaluation with real-world datasets demonstrates superior diagnostic accuracy and resource utilization efficiency, highlighting the transformative potential of IoT and AI-driven approaches in cardiovascular healthcare.