2024/10/04 by Inna, Suma
#ECG #Physical Sciences and Mathematics #artificial neural network #wavelet.
paper · doi:10.17605/osf.io/e5kup
In this paper, the detection of the cardiac abnormalities (abnormalities of heart’s electrical activity) based on recognition of the ECG signal patterns uses two methods, Daubechies wavelet and Artificial Neural Network (ANN). The Daubechies wavelet method is used to reduce the capacity of data without losing the actual information, while ANN is used to recognize ECG patterns. We classified the ECG into normal and abnormal. We made several experiments in training process, only 4 formulas that are used in the identification process. The best accuracy rate is 83% in the formula; input patterns = 240, hidden patterns = 160, output patterns = 3 and the lot of weight = 115200.