Frequency

main_self.py: Use the ecg data collected by ourselves to find the best dimension for classification
main_mit.py: The MIT dataset was used to find the best combination of frequency-domain features suitable for classification
ShowResult.py: Show some results
test.py: Generate the data we want
utils.py: Some functions used in main_self.py、main_mit.py...
text.txt: Experimental results looking for the best combination and the best dimension

Preprocess

rrinterval_seg.m: MIT_BIH dataset segment, based on label_time saved in ATRTIME
Q1.m: Our ecg data preprocess, denoise is not included, which is done via Matlab 1-D Wavelet toolbox.
rdata.m: MIT_BIH dataset preprocess, including how to read the file of 'hea, art...', from Machine_Learning_ECG-master.

Data_H5: FFT data、 DWT data and so on.

data_self_dwt.h5
data_self_fft.h5
data_self_src.h5
......

model_LinearSVC: LinearSVC models trained with FFT and DWT data to find the best dimension.

model_lsvc_fft+dwt_16.pkl
model_lsvc_fft+dwt_32.pkl
model_lsvc_fft+dwt_64.pkl
......

model_pca: PCA models trained with FFT and DWT data which is used as a dimension reduction.

model_pca_fft+dwt_16.pkl
model_pca_fft+dwt_16.pkl
model_pca_fft+dwt_16.pkl
......

out: Frequency domain characteristics of ecg signals.

fear_1_frequency.mat
fear_2_frequency.mat
joy_10_frequency.mat
......