Klasifikasi Sinyal Elektrokardiogram Menggunakan Stockwell Transforms dan K-Nearest Neighbor

Abstrak

An electrocardiogram signal is a bio-electrical signal that results from the electrical activity of the heart. Information on heart health conditions can be inferred by analyzing its shape, rhythm, duration, and orientation. Various methods have been developed to analyze or classify ECG signals automatically. Some of them use the transformation method to convert signals from the time domain to another signal domain. In this study, the Stockwell transform (S-transform) was used to convert signals from the time domain to the time-frequency domain. Minimum and maximum values of the time series of S-transforms were used as K-NN inputs as classifiers. The accuracy of using S-transform was compared with the accuracy of the short-term Fourier transform (STFT), which is an equivalent transformation. The test results showed that S-transform produced higher accuracy compared to FFT on the six classes of ECG signal data tested.