random_search.py

# define the parameter grid
grid_parameters = {'n_estimators': [80, 90, 100, 110, 115, 120],
                   'max_depth': [3, 4, 5, 6],
                   'max_features': [None, 'auto', 'sqrt', 'log2'], 
                   'min_samples_split': [2, 3, 4, 5]}


# define the RandomizedSearchCV class for trying different parameter combinations
random_search = RandomizedSearchCV(estimator=GradientBoostingClassifier(),
                                   param_distributions=grid_parameters,
                                   cv=5,
                                   n_iter=150,
                                   n_jobs=-1)

# fitting the model for random search 
random_search.fit(X_train, y_train)

# print best parameter after tuning
print(random_search.best_params_)