68
правок
Изменения
→Примеры кода: Бустинг
'''Инициализация'''
from pydataset import data
#Считаем данные The Boston Housing Dataset
df = data('Housing')
#Проверим данные
df.head().values
df[i[0]] = df[i[0]].map(d)
df[‘price’] = pd.qcut(df[‘price’], 3, labels=[‘0’, ‘1’, ‘2’]).cat.codes
# Разделим множество на два
y = df['price']
from sklearn.linear_model import RidgeClassifier
from sklearn.svm import SVC
seed = 1075
np.random.seed(seed)
rg = RidgeClassifier()
clf_array = [rf, et, knn, svc, rg]
for clf in clf_array:
vanilla_scores = cross_val_score(clf, X, y, cv=10, n_jobs=-1)
Mean of: 0.632, std: (+/-) 0.081 [RandomForestClassifier]
Mean of: 0.639, std: (+/-) 0.069 [Bagging RandomForestClassifier]
Mean of: 0.636, std: (+/-) 0.080 [ExtraTreesClassifier]
Mean of: 0.654, std: (+/-) 0.073 [Bagging ExtraTreesClassifier]
Mean of: 0.500, std: (+/-) 0.086 [KNeighborsClassifier]
Mean of: 0.535, std: (+/-) 0.111 [Bagging KNeighborsClassifier]
Mean of: 0.465, std: (+/-) 0.085 [SVC]
Mean of: 0.535, std: (+/-) 0.083 [Bagging SVC]
Mean of: 0.639, std: (+/-) 0.050 [RidgeClassifier]
Mean of: 0.597, std: (+/-) 0.045 [Bagging RidgeClassifier]
'''Бустинг'''
ada_boost = AdaBoostClassifier()
grad_boost = GradientBoostingClassifier()
xgb_boost = XGBClassifier()
boost_array = [ada_boost, grad_boost, xgb_boost]
eclf = EnsembleVoteClassifier(clfs=[ada_boost, grad_boost, xgb_boost], voting='hard')
labels = ['Ada Boost', 'Grad Boost', 'XG Boost', 'Ensemble']
for clf, label in zip([ada_boost, grad_boost, xgb_boost, eclf], labels):
scores = cross_val_score(clf, X, y, cv=10, scoring='accuracy')
print("Mean: {0:.3f}, std: (+/-) {1:.3f} [{2}]".format(scores.mean(), scores.std(), label))
# Результат
Mean: 0.641, std: (+/-) 0.082 [Ada Boost]
Mean: 0.654, std: (+/-) 0.113 [Grad Boost]
Mean: 0.663, std: (+/-) 0.101 [XG Boost]
Mean: 0.667, std: (+/-) 0.105 [Ensemble]
== Источники информации ==
