Performance of DL1 considering a fivefold cross-validation
| Predicted 0s | Predicted 1s | |
| True 0s | 47 | 53 |
| True 1s | 11 | 90 |
Sensitivity 89%; specificity 53%.
In contrast to model DL0, this model is excellent at accurately predicting participants (true 1s) but is poor at rejecting inappropriate patients.
DL, deep learning paradigm.