Model | Training AUC | Validation AUC | Test AUC | Performance error | Gerenalisation error |
|---|
ALL | 0.944 | 0.944 | 0.909 | 0.000 | 0.035 |
DM | 0.976 | 0.976 | 0.956 | 0.000 | 0.020 |
DFP | 0.912 | 0.908 | 0.897 | 0.004 | 0.013 |
- The AUCs and error rates from cross-validation for the three SuRFR models. Column 1 shows the three models (ALL, DM, DFP). Columns 2 and 3 show the average training AUCs and validation AUCs, respectively, for each of the three models from the 10-fold cross-validation analysis. The performance error (column 5) shows that the difference between the training and validation AUCs is small. Column 4 shows the average test AUCs achieved by each of the three models run on the hold-out datasets. The low gerenalisation errors in column 6 and the AUCs from the test datasets show that SuRFR is likely to gerenalise and perform equally well on novel data.
