Skip to main content
. 2021 Feb 11;15:618408. doi: 10.3389/fnbot.2021.618408

Table 2.

Classification results of fatigue state vs. alert state for the validation set.

Classification metrics Classification model
ANN-BP ANN-PSO RBF-(ROLS+D-opt) RBF-TLLH
TP 552 517 576 665
TN 554 452 688 670
FN 168 203 144 55
FP 166 268 32 50
Sensitivity/TPR (%) 76.67% 71.81% 80.00% 92.36%
Specificity/TNR (%) 76.94% 62.78% 95.56% 93.06%
Accuracy (%) 76.81% 67.29% 87.78% 92.71%

True represents the fatigue state; false represents the alert state.

TP, true positive; TN, true negative; FN, false negative; FP, false positive.