mlprod:tuning.py

from sklearn.preprocessing import KBinsDiscretizer
from sklearn.model_selection import RandomizedSearchCV

search_space = {
  'preprocessing__binfeatures__n_bins': [10, 20],
  'model__n_estimators': [500, 1000, 2000],
  'model__class_weight': ['balanced', 'balanced_subsample']
}

preprocessing = ColumnTransformer(transformers=[
    ('binfeatures', KBinsDiscretizer(), ['numerical_feature'])
])

pipeline = Pipeline(steps=[
    ('preprocessing', preprocessing),
    ('model', RandomForestClassifier())
])

search = RandomizedSearchCV(pipeline, search_space, cv=5)

search.fit(x, y)

search.best_estimator_