1
$\begingroup$

I have a time series with more than 30 features. For preprocessing with scikit learn do you usually use one scaler per feature or one scaler for all features that should be standardized/normalized?

$\endgroup$
2
$\begingroup$

Sklearn scaler works on feature/column (and thats why you want)

Imagine if it did not. Than you would shift your mean and std in weird-determined-by distribution-of-the-whole-set-kind of way.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy