Table 2.
Results of 10-fold cross-validation for the classification models using the unbalanced and balanced data sets.
| Data sets and models | Precision (%) | Accuracy score (%) | Recall (%) | Brier Score | |||||
| RT-PCRa unbalanced and balanced | |||||||||
|
|
MLPb, unbalanced (balanced) | 97.33 (95.86) | 96.24 (95.81) | 97.08 (95.80) | 0.04 (0.04) | ||||
|
|
GBMc, unbalanced (balanced) | 97.32 (95.95) | 96.30 (95.70) | 97.17 (95.47) | 0.04 (0.04) | ||||
|
|
RFd, unbalanced (balanced) | 97.42 (96.06) | 96.55 (96.00) | 97.46 (95.97) | 0.04 (0.04) | ||||
|
|
DTe, unbalanced (balanced) | 97.49 (96.50) | 96.33 (95.91) | 97.04 (95.32) | 0.04 (0.04) | ||||
|
|
XGBoostf, unbalanced (balanced) | 97.36 (95.94) | 96.30 (95.52) | 97.13 (95.10) | 0.04 (0.04) | ||||
|
|
KNNg, unbalanced (balanced) | 97.38 (95.92) | 96.55 (95.48) | 97.50 (95.04) | 0.03 (0.05) | ||||
|
|
SVMh, unbalanced (balanced) | 97.17 (95.84) | 96.19 (95.58) | 97.18 (95.34) | 0.04 (0.04) | ||||
|
|
LRRi, unbalanced (balanced) | 86.97 (76.86) | 86.72 (81.70) | 94.37 (90.93) | 0.13 (0.18) | ||||
|
|
LRj, unbalanced (balanced) | 87.00 (76.56) | 86.72 (80.63) | 94.33 (88.53) | 0.13 (0.19) | ||||
| Rapid unbalanced and balanced | |||||||||
|
|
MLP, unbalanced (balanced) | 99.33 (96.66) | 98.70 (95.40) | 99.32 (94.10) | 0.01 (0.05) | ||||
|
|
GBM, unbalanced (balanced) | 99.33 (96.18) | 98.72 (95.33) | 99.34 (94.50) | 0.01 (0.05) | ||||
|
|
RF, unbalanced (balanced) | 99.26 (96.42) | 98.76 (95.21) | 99.44 (93.98) | 0.01 (0.05) | ||||
|
|
DT, unbalanced (balanced) | 99.37 (95.51) | 98.69 (94.59) | 99.27 (93.67) | 0.01 (0.05) | ||||
|
|
XGBoost, unbalanced (balanced) | 99.33 (96.83) | 98.72 (95.41) | 99.34 (93.94) | 0.01 (0.05) | ||||
|
|
KNN, unbalanced (balanced) | 99.31 (97.43) | 98.84 (94.58) | 99.49 (91.63) | 0.01 (0.05) | ||||
|
|
SVM, unbalanced (balanced) | 99.30 (97.30) | 98.73 (95.60) | 99.37 (93.85) | 0.01 (0.04) | ||||
|
|
LRR, unbalanced (balanced) | 96.65 (82.00) | 96.23 (84.22) | 99.53 (87.93) | 0.04 (0.16) | ||||
|
|
LR, unbalanced (balanced) | 96.75 (84.75) | 96.14 (85.33) | 99.32 (86.32) | 0.04 (0.15) | ||||
| Both unbalanced and balanced | |||||||||
|
|
MLP, unbalanced (balanced) | 95.36 (93.53) | 94.82 (89.18) | 99.20 (84.23) | 0.05 (0.11) | ||||
|
|
GBM, unbalanced (balanced) | 95.23 (93.67) | 94.73 (89.31) | 99.25 (84.38) | 0.05 (0.11) | ||||
|
|
RF, unbalanced (balanced) | 95.31 (93.81) | 94.87 (89.22) | 99.32 (84.04) | 0.05 (0.11) | ||||
|
|
DT, unbalanced (balanced) | 95.43 (93.75) | 94.79 (89.12) | 99.10 (83.87) | 0.05 (0.11) | ||||
|
|
XGBoost, unbalanced (balanced) | 95.32 (93.60) | 94.78 (89.22) | 99.21 (84.24) | 0.05 (0.11) | ||||
|
|
KNN, unbalanced (balanced) | 95.50 (92.77) | 91.09 (88.63) | 94.81 (83.86) | 0.09 (0.11) | ||||
|
|
SVM, unbalanced (balanced) | 95.21 (93.36) | 94.75 (89.33) | 99.30 (84.73) | 0.05 (0.11) | ||||
|
|
LRR, unbalanced (balanced) | 92.45 (80.79) | 92.04 (80.48) | 99.48 (80.11) | 0.08 (0.20) | ||||
|
|
LR, unbalanced (balanced) | 92.49 (82.44) | 91.98 (81.08) | 99.36 (79.14) | 0.08 (0.19) | ||||
aRT-PCR: reverse transcription polymerase chain reaction.
bMLP: multilayer perceptron.
cGBM: gradient boosting machine.
dRF: random forest.
eDT: decision tree.
fXGBoost: extreme gradient boosting.
gKNN: k-nearest neighbors.
hSVM: support vector machine.
iLRR: logistic regression (strong regularization).
jLR: logistic regression (weak regularization).