Producer risk, statistical validation

False-classification rate with a risk-adjusted cut-off was also considered as a statistic for the cross-validation. The cut-off value corresponds to the fraction of cases in the outcome category. If the predicted risk is greater than the cut-off value, the case is classified as 1 (else 0). The false-classification rate seems to be less suited as a statistic for this study as it is highly sensitive to the low number of cases in each test group and therefore produces high variance. 

---