How to avoid optimal binning of numerical variables while using the transform variables node in Sas

Shradha1 — Mon, 01 Mar 2021 12:21:44 GMT

I am using SAS Enterprise Miner to fit a logistic regression model.

after the Impute node, I am using the transform Variables node with the following instructions:

Interval Inputs= Best

Class Inputs= Dummy Indicators.

After this step, all of my continuous numeric variables are getting transformed into nominal variables being divided into 4 bins for different range of values of the variable. How do I avoid this optimal binning of Numeric variables, so that they remain numeric and do not get transformed to Nominal, whereas for Class variables, dummies are created.

Thank you.

Re: How to avoid optimal binning of numerical variables while using the transform variables node in

MikeStockstill — Wed, 03 Mar 2021 18:29:29 GMT

Hello Shradha1 -

For this data, instead of selecting Best as the default Interval Input method, try one of the other methods. Some transformations create nominal variables (example: Optimal Binning), whereas others create interval variables (example: Log). You can select Multiple if you want to see what the result will be for each transformation, and then you can decide what you want on a per-variable basis.

Have a great day.

topic How to avoid optimal binning of numerical variables while using the transform variables node in Sas in SAS Data Science

How to avoid optimal binning of numerical variables while using the transform variables node in Sas

Re: How to avoid optimal binning of numerical variables while using the transform variables node in