Predicted 0s | Predicted 1s | |
True 0s | 73 | 5 |
True 1s | 8 | 17 |
Sensitivity 93.5%; specificity 68%; positive predictive value (precision) 77%.
At first it appears that there is not a significant improvement obtained via this model but the user can be more certain that the output recommended candidates are more reliable than DL1 or DL0 alone.
DL, deep learning paradigm.