def stacking_gaussian(var_smoothing=1e-9):
estimadores = [('svm', svm()), ('xgboost', xgboost()),
('random_forest', random_forest())]
cv = utils.kfold_for_cross_validation()
stacking = StackingClassifier(
estimators=estimadores,
final_estimator=GaussianNB(var_smoothing=var_smoothing),
stack_method="predict_proba",
cv=cv)
return stacking
# - Se buscan mejores hiperparametros que los default con un GridSearchCV para ambos NB
# #### Hiperparámetros
pipeline_gaussian = Pipeline([("preprocessor", pp.PreprocessingGaussianNB1()),
("model", GaussianNB())
])
pipeline_categorical = Pipeline([("preprocessor", pp.PreprocessingCategoricalNB1()),
("model", CategoricalNB())
])
# +
from sklearn.model_selection import GridSearchCV
params = {'model__alpha': np.arange(1, 10, 1)}
cv = utils.kfold_for_cross_validation()
#Descomentar para ejecutar GridSearchCV
gscv_categorical = GridSearchCV(pipeline_categorical, params, scoring='roc_auc', n_jobs=-1, cv=cv, return_train_score=True).fit(X, y)
print(gscv_categorical.best_score_)
print(gscv_categorical.best_params_)
# +
params = {'model__var_smoothing': [1e-9, 1e-8, 1e-7, 1e-6, 1e-3, 5e-3, 1e-2, 3e-2, 5e-2, 0.1, 0.3]}
cv = utils.kfold_for_cross_validation()
#Descomentar para ejecutar GridSearchCV
gscv_gaussian = GridSearchCV(pipeline_gaussian, params, scoring='roc_auc', n_jobs=-1, cv=cv, return_train_score=True).fit(X, y)
print(gscv_gaussian.best_score_)
print(gscv_gaussian.best_params_)
# -