Skip to main content
. 2021 Feb 25;3:639444. doi: 10.3389/fdgth.2021.639444

Table 4.

Recent publications based on HRV + Machine Learning. The accuracy produced and the theoretical computational cost required by the algorithm.

References Accuracy (%) Computational cost ML algorithm(s)
Castaldo et al. (41) 94,88,94,94 0(n),0(kd),0(nlogn),0(nd2) MLP, SVM, C4.5,LDA
Cho et al. (70) 90.19 0(n·k·d) CNN
Cho et al. (26) 95 0(n4) K-ELM
Coutts et al. (71) 83 0(W)
W = 4IH+4H2+3H+HK
LTSM
Taye et al. (72) 98.6 0(W)
W = IH+HK
ANN
Arsalan et al. (73) 92.85 0(n) MLP
Lima et al. (38) 80 0(n*log(n)*d*k) Random Forest
Kublanov et al. (74) 91.3,87.8,
87.1,88.2
0(nd2),0(kd),
0(n*log(n)*d), 0(c*d)
LDA,SVM,DT,NB
Ma et al. (75) 96.58,
98.2
0(n·k·d),
0(n)
CNN,
MLP
Persson et al. (76) 77.5,83.4,
82.4,85.4
0(nd),0(n2),
0(nt),0(n*log(n)*d*k)
KNN, SVM,
AdaBoost, RF