Table 6 Comparison with other methods

Ebrahimzadeh et al. [9]

ECG signals

Linear (time, frequency domain), TF domain, and nonlinear methods (Poincare, detrended fluctuation analysis)

MLP

12 min:

Acc = 83.88%

Sen = 82.67%

Spe = 85.09%

Lopez-Caracheo et al. [10]

ECG signals

Nonlinear methods (Higuchi fractal dimension, Box dimension, Katz fractal dimension)

MLP-NN

14 min:

Acc = 91.40%

Heng et al. [14]

HRV signals

Linear (time domain) and nonlinear methods (Hurst Exponent, SD1)

SVM

4 min:

Acc = 94.70%

Sen = 100%

Spe = 88.90%

Ebrahimzadeh et al. [15]

HRV signals

Linear (time, frequency domain), TF and nonlinear methods (Poincare and detrended fluctuation analysis)

KNN, SVM, ME, and MLP

13 min:

Acc = 84.28%

Sen = 85.72%

Spe = 82.86%

Shi et al. [16]

HRV signals

EEMD, linear (time, frequency domain), TF domain, and nonlinear methods (Rényi entropy, fuzzy entropy, dispersion entropy, Rényi distribution entropy and improved multiscale permutation entropy)

KNN

14 min:

Acc = 96.1%

Sen = 97.5%

Spe = 94.4%

Ours

HRV signals

Linear (SDRR) and nonlinear methods (Shannon entropy, \(S_{v}\))

SVM

Acc for 5, 20, 35, 60 min:

95.00%, 94.29%, 97.50%, 92.50%

Average:

Acc = 91.22%

Sen = 96.15%

Spe = 89.59%