Discretize `Table`

[lars-reimann]

Is your feature request related to a problem?

Discretization means to replace a continuous variable by a variable that only has a finite amount of values. This is a preprocessing step that we should support.

Desired solution

• Add a class Discretizer in safeds.data.tabular.transformation that wraps the KBinsDiscretizer of scikit-learn
• Make the class a subclass of TableTransformer
• The __init__ should for now only have a parameter number_of_bins to control how many bins are created
• If number_of_bins is less than 2, raise a ValueError

[guenterk]

@robmeth : Please add a comment explaining the problem why you marked this as blocked (you mentioned in the final stand up meeting today the failing pandasEqualsTest...).

[Marsmaennchen221]

@robmeth Use the ordinal encoding. This transforms the data and returns the bin index rather than a bin as sparse matrix. This will also resolve the problem with the tests. See #327 (comment)