Update CategoricalTransformer to make it user friendly

[npatki]

Problem Description

Let's update CategoricalTransformer to make it more user friendly.

Expected behavior

Name: FrequencyEncoder
Functionality:

• Change columns (see Rename columns parameter in fit for each individual transformer #376 )
• Change fuzzy to add_noise
• Remove clip
• Add warning for unseen categories

>>> from transformers.categorical import FrequencyEncoder
# (default) No noise
>>> fe = FrequencyEncoder() 

>>> fe = FrequencyEncoder(add_noise=True)
>>> fe.fit(data, column='categorical_column')

# if the transform contains new categories unseen during fit
>>> fe.transform(data_new_categories)
Warning: The data contains new categories ('hello', 'world') that were not in the original data.
Assigning them random values. If you want to model new categories, please fit the transformer
again with the new data.

# reverse transform: always clip without warning (ie never use modulo)
>>> fe.reverse_transform(model_data)