Update names & functionality for handling missing values

[npatki]

Problem Description

Let's update the usage of missing value handling to make it more user friendly. The update would apply to all transformers that handle missing values. Currently, these are:

• BooleanTransformer
• DateTimeTransformer (+ variants)
• NumericalTransformer (+ variants)
• NullTransformer

Expected behavior

1. nan becomes missing_value_replacement, with updated defaults

# (default) do not replace missing values
nt = NumericalTransformer()

# replace with a static number
nt = NumericalTransformer(missing_value_replacement=0.00)

# replace using a method
nt = NumericalTransformer(missing_value_replacement='mode')

2. null_column becomes model_missing_values that can only accept True/False values; updated functionality & defaults

# (default = False) do not create a new column
nt = NumericalTransformer()
nt = NumericalTransformer(model_missing_values=False)

# True creates a missing column if there is missing data
nt = NumericalTransformer(model_missing_values=True)
nt.fit(data, columns=['column_with_missing_values'])

# otherwise do not create a column and warn the user
nt = NumericalTransformer(model_missing_values=True)
nt.fit(data, columns=['never_null_column'])
Guidance: There are no missing values in column 'never_null_column'. Extra column not created.

[amontanez24]

A couple points:

1. This issue will probably have to be completed at the same as Update NullTransformer to make it user friendly #372 since all these transformers call the NullTransformer behind the scenes, so they will crash if the parameters for the NullTransformer change and theirs don't
2. Since each transformer will be using the same parameters to define how the NullTransformer will be used, perhaps it makes sense to move them to the BaseTransformer class. On top of that, we may want to call the NullTransformer from within the BaseTransformer's fit, transform and reverse_transform methods. Something like

def fit(self, data, columns):
    ...
    self.null_Transformer = NullTransformer(missing_value_replacement=self.missing_value_replacement, model_missing_values = self.model_missing_values)
    self._fit(columns_data)
    self.null_transformer.fit(columns_data)
    
def transform(self, data):
    ...
    transformed_data = self._transform(columns_data)
    transformed_data = self.null_transformer.transform(data)

def reverse_transform(self, data):
    ...
    columns_data = self.null_transformer.reverse_transform(columns_data)
    reversed_data = self._reverse_transform(columns_data)

[pvk-developer]

Closed by #384