feat: Baseline Model Training

[zogomii]

Is your feature request related to a problem?

Subtask of #710

Desired solution

Create method Baseline._model(table:TabularDataSet, metric:Metric, positiveClass: List)->Model

Tasks:

1. split data - 70% train data and 30% test data
2. control if there are more than one values in positiveClass. If yes, calculate the mean of every positive class, else just for one positive Class (?)
3. if there's more than 2 unique values in target column calculate the mean of every positive class, else require the positive class input and calculate its metric (?)
4. for-loop for every regressor or classifier
5. choosing the best (just gradient boosting, random forest and linear regressor with simple standard-values)

Possible alternatives (optional)

No response

Screenshots (optional)

No response

Additional Context (optional)

No response

[lars-reimann]

AutoML is currently out-of-scope.