Electrodiogram Signal Classification by Using XGBoost in Different Discrete Wavelet Transform

Aji, Bibit Waluyo (Department of Mathematics, Universitas Diponegoro, Semarang)
Chasanah, Sri Nur (Department of Mathematics, Universitas Diponegoro, Semarang)
Sa’adah, Fihris Aulia (Department of Nursing, Universitas Diponegoro, Semarang)
Irawanto, Bambang (Department of Mathematics, Universitas Diponegoro, Semarang)

International Journal of Computing Science and Applied Mathematics
Vol 10, No 2 (2024)

Electrocardiogram (ECG) is electrical signal from heart. ECG can use for Detection or tracking the hearth health. The one of method can use is machine learning. Machine learning is Algorithm which can learning from data and is used for classifying and predicting. Machine Learning can use for signal classification, in this case is for ECG classification. In signal processing, wavelet transform is common method for analyzing signal. Wavelet transform has many familly. The aim from this research is to find the best wavelet transform in the classification of Electrocardiogram (ECG) signals on XGBoost. The Discrete Wavelet Transform which is used for the research is daubechies, coiflets, symlets, biorthogonal, reverse biorthogonal, haar. Finally, the best wavelet transform in the classification is biorthogonal (3.1) with F1 score 1.0.