Detection of Sleep Apnea from Single-Lead ECG Signal Using a Time Window Artificial Neural Network

Table 2

Performance of TW-MLP and traditional non-time window machine learning methods on per-recording classification.

Method	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC	Corr.
LR	91.4	100.0	75.0	1.000	0.850
LDA	88.6	100.0	66.7	1.000	0.855
SVM	82.9	100.0	50.0	0.975	0.875
MLP	82.9	100.0	50.0	0.986	0.851
TW-MLP	97.1	100.0	91.7	1.000	0.935