Automated feature removal

[karanmohan357]

I can't remove the features which are auto generated during training.I tried many methods but the features are still auto generating entirely.

Steps to reproduce the behavior:

1. go to train_model_pipeline and train the model
2. Features are automatically generated even if I mention the features to use in predictionjobclass

To be able to generate the feature we want during training. Instead of auto generating as some features are not relevant in another timezone

[karanmohan357]

To select which feature to autogenerated as some features are irrelevant in another timezone

[egordm]

Hi, yes unfortunately that's a known limitation in openstef 3.
Workarounds known to me:

• Adding custom feature preprocessor. See prediction job, one should be able to pass a custom class there. (https://github.com/OpenSTEF/openstef/blob/main/openstef/data_classes/prediction_job.py#L144)
• skip model train pipeline and just do the preprocessing and use the inner model directly.

We are actually currently working on openstef v4 where this issue is actually fixed.
There you can control the whole feature pipeline yourself. See Openstef 4 model presets.
You can install it using pip install openstef==4.0.0.a10. Depending on your use case you might actually want to prefer using it.

Let me know if it works for you.