[DOC] Enhance BestEstimator docstrings

[SimonBlanke]

File: src/hyperactive/integrations/sklearn/best_estimator.py

Description: The BestEstimator class has minimal one-word docstrings. This should be improved.

Methods to document:

• score_samples()
• predict()
• predict_proba()
• decision_function()
• transform()
• inverse_transform()
• classes_

What to do:

• Add parameter descriptions (X: array-like)
• Add return type documentation
• Add better description