A machine learning model is calibrated if its predicted probability for an
outcome matches the observed frequency for that outcome conditional on the
model prediction. This property has become increasingly important as the impact
of machine learning models has continued to spread to va