ECG cardiac abnormality signal classification using HMLP network

Since irregular heartbeat symptoms arise, it's critical to accurately diagnose the patient's heart problem. This research aims to use a training algorithm for disease detection. The amplitude and duration of P peaks, QRS waves and T peaks are among the data of ficidual detection points that the electrocardiogram (ECG) signal offers. These informational points serve as input parameters. To further develop the prediction model, the Hybrid Multilayer Perceptron (HMLP) network was employed to perform the prediction. To estimate the correctness of the ECG signal model more accurately, Bayesian regularization (BR) training algorithm is applied. HMLP network trained by BR training algorithm capable of performing with mean squared error (MSE) of 0.32 and a regression value of 0.96. The trained model uses Tansig activation function to activate the network.