Discrete mathematical model for heart disease ECG waveform using Kernel function

Abstract

In this paper proposed a new design of the mathematical model for ECG waveform using Kernel function. The continued standard ECG signal is converted to the discrete time signal or original ECG data. After that the original ECG data is changed to the Kernel ECG data by Kernel function. Then the Kernel ECG data is converted to mathematical model using discrete least square technique. At the time this will be the complete mathematical model of ECG equation and will be used to implement the ECG simulator with high resolution and low memory storage. For the accuracy of the ECG signals that are implemented by mathematical models with ECG simulations is presented in a PRD (Percent root mean square difference) and MAPE ( Mean absolute percentage error). In this stud, six kernel functions were compared on PRD and MAPE. The kernel functions were (1) Epanechnikov, (2) Quartic, (3) Triweight, (4) Triangular, (5) Gaussian and (6) Cosin. Each function was subjected to three difference kind of heart disease. The result from these studies showed that the Kernel Triweight Function lowest  MAPE and PRD when  compared to the other Kernel Function