CODEWITHSUNDEEP
pythonautoml
Rosa Aghdam
Rosa Aghdam

Reputation: 131

skipping pre-processing step in H2OAutoML

I apply H2OAutoML for classification purposes. In the first step, I developed several normalization techniques for my data. Then, I would want to apply H2OAutoML to MY normalized datasets and compare the results depending on various normalizing techniques. But, H2OAutoML first normalized the data and then ran the model on it; I would want to skip the normalization step in H2OAutoML, but I cannot set the option. I'm curious if it's possible to deactivate the normalization phase and evaluate my data.

For instance, the standardize option is available in H2OGeneralizedLinearEstimator(standardize = False), but not in H2oautoML (H2OAutoML received an unusual keyword argument'standardize').

Thanks for your help and time!

Upvotes: 0

Views: 57

Answers (1)

Rosa Aghdam
Rosa Aghdam

Reputation: 131

Based on this R help "https://www.rdocumentation.org/packages/h2o/versions/3.36.1.2/topics/h2o.automl" the preprocessing option is NULL by default (preprocessing = NULL). Therefore, my data is not normalized by H2OAutoML.

Upvotes: 1

Related Questions