Prior Probabilities / Prevalences

Use this feature to specify the assumed prior proportions of each of the classes.

Note : Custom is not available for all processes. If you choose Custom , you must enter the assumed prior probabilities for each of the classes in the Custom Prior Probabilities text field.

Your choices are described below:

Prior Probabilities / Prevalences	Definition
Equal	Choose Equal when all of the groups have the same prior probability . Note : When you choose this option, the performance metrics ( Root Mean Square Error , MAE, and Accuracy) are adjusted to assign equal prevalences to the classes.
Proportional	Choose Proportional when the prior probabilities of the groups are equal to those of the input data set.
Custom	Choose Custom when you want to specify your own prior probabilities / prevalences. This is useful in cases where you have prior knowledge about what they are and want them to influence the posterior probabilities.

To Specify the Prior Probabilities / Prevalences:

Choose Automated as the mode for generating the model (if available) .

Select the desired option by clicking within the appropriate radio button.