Figure 7
Calibration curves for the training, validation, and test sets using MC dropout (a) and direct probability prediction (b). (c) and (d) show the difference between the observed and expected cumulative probability using MC dropout and direct probability prediction, respectively.
