Table 3.
Prediction performance and variables selected for each model for the five models with a cvAUC >0.75
| Model | cvAUC | Activity score | Chronicity score |
Interstitial fibrosis | Interstitial inflammation | Urine protein-to-creatinine ratio | WBC | Hgb | # of predictors |
| LR | 0.780 | X | X | X | X | X | 5 | ||
| RF | 0.800 | X | X | X | X | 4 | |||
| SVML | 0.783 | X | X | X | X | 4 | |||
| SVMR | 0.790 | X | X | X | X | X | X | 6 | |
| ANN | 0.775 | X | X | X | X | X | 5 |
ANN, artificial neural networks; AUC, area under the curve; cvAUC, cross-validated AUC; Hgb, haemoglobin; LR, logistic regression; RF, random forest; WBC, white blood cell.