Table 2 Training and test set performance of the best models. For the training set the mean of each fold ± 1SD is reported
Performance Metrics | Genotype model, mean ± 1SD (test) | Phenotype model, mean ± 1SD (test) | Combined model, mean ± 1SD (test) | Additive model, mean ± 1SD (test) |
|---|---|---|---|---|
AUC | 0.53 ± 0.06 (0.56) | 0.63 ± 0.05 (0.68) | 0.68 ± 0.06 (0.72) | 0.58 ± 0.03 (0.67) |
Balanced accuracy | 0.53 ± 0.04 (0.54) | 0.59 ± 0.04 (0.65) | 0.64 ± 0.05 (0.64) | 0.56 ± 0.03 (0.63) |
Sensitivity | 0.59 ± 0.03 (0.55) | 0.60 ± 0.02 (0.62) | 0.63 ± 0.02 (0.70) | 0.69 ± 0.01 (0.70) |
Specificity | 0.60 ± 0.03 (0.55) | 0.60 ± 0.02 (0.62) | 0.63 ± 0.02 (0.71) | 0.61 ± 0.02 (0.71) |