1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22

from sklearn.model_selection import GridSearchCV

  

param_grid = [
# try 12 (3×4) combinations of hyperparameters
{'n_estimators': [3, 10, 30], 'max_features': [2, 4, 6, 8]},

# then try 6 (2×3) combinations with bootstrap set as False
{'bootstrap': [False], 'n_estimators': [3, 10], 'max_features': [2, 3, 4]}
]

  

forest_reg = RandomForestRegressor(random_state=42)

# train across 5 folds, that's a total of (12+6)*5=90 rounds of training
grid_search = GridSearchCV(forest_reg, param_grid, cv=5,
			scoring='neg_mean_squared_error',
			return_train_score=True)

grid_search.fit(housing_prepared, housing_labels)

1
2
3
4

cvres = grid_search.cv_results_

for mean_score, params in zip(cvres["mean_test_score"], cvres["params"]):
	print(np.sqrt(-mean_score), params)

1
2

>>> 
65005.182970763315 {'max_features': 2, 'n_estimators': 3} 55582.91015494046 {'max_features': 2, 'n_estimators': 10} 52745.33887865031 {'max_features': 2, 'n_estimators': 30} 60451.18914812725 {'max_features': 4, 'n_estimators': 3} 53062.818497303946 {'max_features': 4, 'n_estimators': 10} 50663.79774079741 {'max_features': 4, 'n_estimators': 30} 57998.07162873506 {'max_features': 6, 'n_estimators': 3} 52042.04702364244 {'max_features': 6, 'n_estimators': 10} 50028.060190761295 {'max_features': 6, 'n_estimators': 30} 58308.44501796401 {'max_features': 8, 'n_estimators': 3} 52082.74313186547 {'max_features': 8, 'n_estimators': 10} 50165.81805010987 {'max_features': 8, 'n_estimators': 30} 62709.54311517104 {'bootstrap': False, 'max_features': 2, 'n_estimators': 3} 54062.01766032325 {'bootstrap': False, 'max_features': 2, 'n_estimators': 10} 60613.541905953585 {'bootstrap': False, 'max_features': 3, 'n_estimators': 3} 53742.988651846914 {'bootstrap': False, 'max_features': 3, 'n_estimators': 10} 59387.46561811065 {'bootstrap': False, 'max_features': 4, 'n_estimators': 3} 52826.41762121993 {'bootstrap': False, 'max_features': 4, 'n_estimators': 10}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15

from sklearn.model_selection import RandomizedSearchCV
from scipy.stats import randint

param_distribs = {
'n_estimators': randint(low=1, high=200),
'max_features': randint(low=1, high=8),
}

forest_reg = RandomForestRegressor(random_state=42)

rnd_search = RandomizedSearchCV(forest_reg, param_distributions = param_distribs, n_iter=10, cv=5,
				scoring='neg_mean_squared_error',
				random_state=42)

rnd_search.fit(housing_prepared, housing_labels)

1
2
3
4
5
6
7

cvres = rnd_search.cv_results_

for mean_score, params in zip(cvres["mean_test_score"], cvres["params"]):
	print(np.sqrt(-mean_score), params)

>>>
49462.596134607906 {'max_features': 7, 'n_estimators': 180} 51676.97211565583 {'max_features': 5, 'n_estimators': 15} 50827.83871022729 {'max_features': 3, 'n_estimators': 72} 51117.698297994146 {'max_features': 5, 'n_estimators': 21} 49585.185219390754 {'max_features': 7, 'n_estimators': 122} 50836.040148806715 {'max_features': 3, 'n_estimators': 75} 50746.890270152086 {'max_features': 3, 'n_estimators': 88} 49788.190631507045 {'max_features': 5, 'n_estimators': 100} 50574.565725719985 {'max_features': 3, 'n_estimators': 150} 65153.787556165735 {'max_features': 5, 'n_estimators': 2}

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

final_model = grid_search.best_estimator_

X_test = strat_test_set.drop("median_house_value", axis=1)
y_test = strat_test_set["median_house_value"].copy()
X_test_prepared = full_pipeline.transform(X_test)

final_predictions = final_model.predict(X_test_prepared)

final_mse = mean_squared_error(y_test, final_predictions)
final_rmse = np.sqrt(final_mse)

1
2

final_rmse
>> 47362.98158022501

1
2
3
4
5
6

from scipy import stats

confidence = 0.95
squared_errors = (final_predictions - y_test) ** 2
np.sqrt(stats.t.interval(confidence, len(squared_errors) - 1,loc=squared_errors.mean(),
scale=stats.sem(squared_errors)))

1

>>> array([45397.61151846, 49249.98392646])