mlr3automl with time series data?

[GitHubGeniusOverlord]

Hi!
Is there at this moment an option to use mlr3automl for data where time is important?
Most importantly it should use a resampling that respects time. However, as far as I know, the mlr3temporal package does not yet provide these options.
Is there a way to do temporal train- test split with mlr3automl?

Thank You!

[a-hanf]

Hi,
there's currently no support for time series data in mlr3automl. I think the following changes would be needed:

• exposing the method for the outer resampling as a parameter in the AutoMLBase$resample
• adjusting the preprocessing pipeline (e. g. removing the imputation, as it might leak information from future dates).
• adding default learners for time series data from mlr3temporal package, ideally in a separate class (similar to AutoMLClassif.R or the regression equivalent)
• adding sensible default hyperparameters to tune over (in get_default_params)

I don't actively develop this package these days, but I'm happy to review a PR if you want to try