Ml process v2

[MichaelBrueggemann]

Features:

• added "train_model" process to train a Random Forest Model with fixed hyperparameters
• added "predict_model" process to make a classification based on a pre-trained model
• added support for exporting data.frames as NetCDF (rough draft, better type checking of datacubes is needed)
• added direct saving of models to the 'user workspace'
  • would be better if this behavior is integrated in "save_results()"