If the AUC thresholds follow a normal distribution (maybeNormalDistr returns TRUE) the threshold is calculated as outliers of the global distrubtion according to the thrP value set by the user. However, the current formula is not given the expected results, since thrP is divided by nCells (see line below). Instead it should be qnorm(1-thrp,mean=meanAUC, sd=sdAUC), or am I mistaken?
If the AUC thresholds follow a normal distribution (maybeNormalDistr returns TRUE) the threshold is calculated as outliers of the global distrubtion according to the thrP value set by the user. However, the current formula is not given the expected results, since thrP is divided by nCells (see line below). Instead it should be qnorm(1-thrp,mean=meanAUC, sd=sdAUC), or am I mistaken?
aucThrs["outlierOfGlobal"] <- qnorm(1-(thrP/nCells), mean=meanAUC, sd=sdAUC)