Abstract:Background: Heart disease and other cardiovascular conditions remain a significant global health challenge, contributing to a high number of fatalities annually. Despite technological advancements, current cardiovascular monitoring methods, especially those analyzing electrocardiograms (ECGs), often fail to capture intricate patterns and subtle irregularities, highlighting the need for a more sophisticated approach. This study proposes a novel framework utilizing deep learning, specifically combining the radial basis function (RBF) for feature extraction and a deep belief network (DBN) for classification, to enhance ECG data analysis. Methods: The proposed method involves preprocessing ECG signals to reduce noise, correct baseline drift, and scale amplitude. Feature extraction is performed using the RBF, which captures intricate temporal patterns in the ECG signals. Subsequently, the DBN classifies the extracted features, leveraging its hierarchical learning capabilities to identify subtle correlations and patterns. The model's performance was evaluated using Python simulations on a high-performance computing system, benchmarked against existing methods including Dual-Stage DL, DWT-ML, and M1M2. Results: The RBF-DBN model demonstrated superior performance across several metrics, achieving 99% accuracy, precision, and recall at the 1000th time step, outperforming Dual-Stage DL, DWT-ML, and M1M2 methods. The RBF-DBN method demonstrated a remarkable accuracy of 99%, surpassing the current Dual-Stage DL method by 2%. Additionally, the approach maintained exceptional sensitivity and recall at 99%. The precision of the model was also 99%, and the F1 score, which balances recall and precision, further underscored the model's efficacy in real-time cardiovascular monitoring. Conclusion: The integration of RBF and DBN in ECG analysis significantly enhances the precision and accuracy of cardiovascular monitoring.

The Deep Radial Basis Function Data Descriptor (D-RBFDD) Network: A One-Class Neural Network for Anomaly Detection

Deep-RBF Networks for Anomaly Detection in Automotive Cyber-Physical Systems

Enhanced Cardiovascular Monitoring Using Radial Basis Function and Deep Belief Network Fusion

Anomaly Detection using One-Class Neural Networks

Deep one-class classification model assisted by radius constraint for anomaly detection of industrial control systems

Novel Data Classification Method Based on Radial Basis Function Networks

A Complete Deep Support Vector Data Description for One Class Learning.

Deep Anomaly Detection by Residual Adaptation

DDR-ID: Dual Deep Reconstruction Networks Based Image Decomposition for Anomaly Detection

DASVDD: Deep Autoencoding Support Vector Data Descriptor for Anomaly Detection

Deep residual shrinkage relation network for anomaly detection of rotating machines

DL-Based Anomaly Detection at the Physical-Layer of Cognitive Radio by Deep Support Vector Data Description

Dual-Branch Reconstruction Network for Industrial Anomaly Detection with RGB-D Data

Deep-Anomaly: Fully Convolutional Neural Network for Fast Anomaly Detection in Crowded Scenes

Deep learning based anomaly detection in real-time video

Learning in Deep Radial Basis Function Networks

Multiresolution dendritic cell algorithm for network anomaly detection

Deep Dual Support Vector Data Description for Anomaly Detection on Attributed Networks

Deep RAN: A Scalable Data-driven platform to Detect Anomalies in Live Cellular Network Using Recurrent Convolutional Neural Network

AnoDFDNet: A Deep Feature Difference Network for Anomaly Detection

Multi-Class Classification of Abnormal Heartbeat Detection using Hyperdimensional Computing