H2O Generalized Linear model learner cannot find target column.

Hey,

I am trying to use the H2O GLM regression learner to learn a lasso regression model. I have succesfully transformed my data into a H2O table with 3 columns:

  1. the Row IDs
  2. Temperature (D)
  3. Max load (I)

Up untill this point there are no errors, but the H2O GLM learner returns this error:

Selected target column null can not be found in input frame. Please select a valid target column in dialog.

When I try to open the GLM learner node, I get the following error:

The dialog cannot be opened for the following reason: No column in spec compatible to “NominalValue”.

I have looked at the 03_H2O_GLM_Regression_Model and my structure is similar to that one. The only difference I can think of that causes this error is in the H2O table. The table in the example has 4 columns with (D) values, mine has one with (D) and on with (I). I have tried normalizing the data, but no luck there either.

Could someone help me with this problem?

Hi @s156284,

Since you want to learn a regression model, I guess you need to use the H2O Generalized Linear Model Learner (Regression) . It accepts numerical columns as target. The non-regression learner, i.e. classification, expects a nominal column. If you have a target with classes encodes as integers, you would need to use the Number to String node to change the type of the column.

Cheers,
Simon

6 Likes

Thank you very much Simon! Somehow I completely missed that I was using the wrong node.

3 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.