bank_marketing_grid_search_1.py

# Create first pipeline for base without reducing features.

pipe = Pipeline([('classifier' , RandomForestClassifier())])
# pipe = Pipeline([('classifier', RandomForestClassifier())])

# Create param grid.

param_grid = [
    {'classifier' : [LogisticRegression()],
     'classifier__penalty' : ['l1', 'l2'],
    'classifier__C' : np.logspace(-4, 4, 20),
    'classifier__solver' : ['liblinear']},
    {'classifier' : [RandomForestClassifier()],
    'classifier__n_estimators' : list(range(10,101,10)),
    'classifier__max_features' : list(range(6,32,5))}
]

# Create grid search object

clf = GridSearchCV(pipe, param_grid = param_grid, cv = 5, verbose=True, n_jobs=-1)

# Fit on data

best_clf = clf.fit(X_train, y_train)