Options

Use this drop-down menu to select between the various classification algorithms available in this module.

 

Binning is the method used to build the linear splines. The auto binning feature is recommended, but takes more time than specifiying the number of bins.

 

Auto Building will iteratively build models by sequentially adding all the variables and then removing unimportant ones. The final model will only contain variables that improve the model by more than the amount specified in the 'minimum gini increase'.

This can be a relatively slow process but is recommended for the final model build.

 

In the Scorecard Builder there is the option to use the specified sup-population as the validation set.