[Feature] Hyperparameter optimization pipeline compatible with new sklearn models

[JanMaartenvanDoorn]

WHAT: With the new pipeline based on sklearn model updating the hyperparameter optimization pipeline is necessary. Here we implement a new pipeline that works with the new sklearn models.

HOW:

• have a look at the examples here: https://scikit-learn.org/stable/modules/grid_search.html
• Focus here on the parameter tuning and try to keep the pipeline lean
• Direct DB calls from the pipeline are undesired, training data should be an input argument
• Also have a look at the method we used with the old pipeline: https://github.com/alliander-opensource/short-term-forecasting/blob/develop/openstf/pipeline/optimize_hyperparameters.py
• Make a plan how to implement the new pipeline
• Featureset optimization is a nice to too have but not strictly necessary, make a new story if you skip this.
• Implement it

TEST:

• hyperparameters of the sklearn models can be tuned automatically with a specialized pipeline

[FrankKr]

added link to old pipeline: https://github.com/alliander-opensource/short-term-forecasting/blob/develop/openstf/pipeline/optimize_hyperparameters.py