Showing results for 
Search instead for 
Did you mean: 

Why DataRobot use One-hot and Ordinal encoding together?

Why DataRobot use One-hot and Ordinal encoding together?

I trained a model and the blueprint looks like this. I'm trying to understand what are the data processing approach here.

Does DR take all my categorical variables through both ordinal and one-hot encoding, then combine them together? What's the benefit of encoding twice? 

Or does it take some categorical variables through ordinal and others through one-hot? If so, where can I find which variables?

And what does the ridge regressor do here?

Any insights are appreciated! Thank you


Labels (1)
0 Kudos
1 Reply
DataRobot Alumni

Hello Ovacation - thanks for reaching out!

Let me see if I can give you some pointers to content that may help.

  • Does your dataset include categorical features? This community article may help (also see the attached PPT).
  • Are you running the trial? If yes, then have a look at the in-app Platform Documentation here for good info that should help (that link will only work for the trial, app2 instance).

Please let me know if you are looking for something more - I'm sure other DS community members will be able to help you!

- Linda

0 Kudos