13-1什么是集成学习
Notbook 示例
Notbook 源码
集成学习
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
[4]
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
[5]
from sklearn.linear_model import LogisticRegression
log_clf = LogisticRegression()
log_clf.fit(X_train, y_train)
log_clf.score(X_test, y_test)
0.864
[6]
from sklearn.svm import SVC
svm_clf = SVC()
svm_clf.fit(X_train, y_train)
svm_clf.score(X_test, y_test)
0.896
[7]
from sklearn.tree import DecisionTreeClassifier
dt_clf = DecisionTreeClassifier()
dt_clf.fit(X_train, y_train)
dt_clf.score(X_test, y_test)
0.84
[8]
y_predict1 = log_clf.predict(X_test)
y_predict2 = svm_clf.predict(X_test)
y_predict3 = dt_clf.predict(X_test)
[9]
y_predict = np.array( (y_predict1 + y_predict2 + y_predict3) >= 2, dtype='int')
[10]
y_predict[:10]
array([1, 0, 0, 1, 1, 1, 0, 0, 0, 0])
[11]
from sklearn.metrics import accuracy_score
accuracy_score(y_test, y_predict)
0.904
Voting Classifier
[12]
from sklearn.ensemble import VotingClassifier
voting_clf = VotingClassifier(estimators=[
('log_clf', LogisticRegression()),
('svm_clf',SVC()),
('dt_clf', DecisionTreeClassifier())
],voting='hard')
[13]
voting_clf.fit(X_train, y_train)
VotingClassifier(estimators=[('log_clf', LogisticRegression()),
('svm_clf', SVC()),
('dt_clf', DecisionTreeClassifier())])
[14]
voting_clf.score(X_test,y_test)
0.912
13-2 SoftVoting Classifier
Notbook 示例
Notbook 源码
Soft Voting
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
[4]
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
Hard Voting Classifier
[5]
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import VotingClassifier
voting_clf = VotingClassifier(estimators=[
('log_clf', LogisticRegression()),
('svm_clf',SVC()),
('dt_clf', DecisionTreeClassifier())
],voting='hard')
[6]
voting_clf.fit(X_train, y_train)
voting_clf.score(X_test,y_test)
0.912
Soft Voting Classifier
[7]
voting_clf2 = VotingClassifier(estimators=[
('log_clf', LogisticRegression()),
('svm_clf',SVC(probability=True)),
('dt_clf', DecisionTreeClassifier())
],voting='soft')
[8]
voting_clf2.fit(X_train, y_train)
voting_clf2.score(X_test,y_test)
0.92
13-3 Bagging和Pasting
Notbook 示例
Notbook 源码
Bagging和Pasting
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
[4]
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
使用 Bagging
[5]
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import BaggingClassifier
bagging_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=100,
bootstrap=True)
[6]
%%time
bagging_clf.fit(X_train, y_train)
bagging_clf.score(X_test,y_test)
CPU times: total: 1.47 s
Wall time: 1.51 s
0.904
[7]
bagging_clf2 = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=5000,max_samples=100,
bootstrap=True)
[8]
%%time
bagging_clf2.fit(X_train, y_train)
bagging_clf2.score(X_test,y_test)
CPU times: total: 15 s
Wall time: 15.2 s
0.912
13-4 oob（Out-of-Bag）和关于Bagging的更多讨论
Notbook 示例
Notbook 源码
obb 和更多Bagging相关
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
使用obb
[4]
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import BaggingClassifier
bagging_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=100,
bootstrap=True, oob_score=True)
bagging_clf.fit(X,y)
bagging_clf.oob_score_
0.92
n_jobs
[5]
%%time
bagging_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=100,
bootstrap=True, oob_score=True )
bagging_clf.fit(X,y)
bagging_clf.oob_score_
CPU times: total: 1.77 s
Wall time: 1.81 s
0.918
[6]
%%time
bagging_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=100,
bootstrap=True, oob_score=True,
n_jobs =-1 )
bagging_clf.fit(X,y)
bagging_clf.oob_score_
CPU times: total: 453 ms
Wall time: 6.93 s
0.918
bootstrap_features
[7]
%%time
random_subspaces_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=500,
bootstrap=True, oob_score=True,
n_jobs =-1 ,
max_features=1, bootstrap_features=True)
random_subspaces_clf.fit(X,y)
random_subspaces_clf.oob_score_
CPU times: total: 438 ms
Wall time: 1.38 s
0.82
[8]
%%time
random_patches_clf = BaggingClassifier(DecisionTreeClassifier(),
n_estimators=500,max_samples=100,
bootstrap=True, oob_score=True,
n_jobs =-1 ,
max_features=1, bootstrap_features=True)
random_patches_clf.fit(X,y)
random_patches_clf.oob_score_
CPU times: total: 469 ms
Wall time: 1.31 s
0.856
13-5 随机森林和Extra-Trees
Notbook 示例
Notbook 源码
随机森林和Extra-Trees
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
随机森林
[4]
from sklearn.ensemble import RandomForestClassifier
rf_clf = RandomForestClassifier(n_estimators=500,random_state=666,
oob_score=True, n_jobs=-1)
rf_clf.fit(X, y)
RandomForestClassifier(n_estimators=500, n_jobs=-1, oob_score=True,
random_state=666)
[5]
rf_clf.oob_score_
0.896
[6]
rf_clf2 = RandomForestClassifier(n_estimators=500,max_leaf_nodes=16,
random_state=666,
oob_score=True, n_jobs=-1)
rf_clf2.fit(X, y)
rf_clf2.oob_score_
0.92
使用 Extra-Trees
[7]
from sklearn.ensemble import ExtraTreesClassifier
et_clf = ExtraTreesClassifier(n_estimators=500, bootstrap=True,
oob_score=True,random_state=666)
et_clf.fit(X, y)
ExtraTreesClassifier(bootstrap=True, n_estimators=500, oob_score=True,
random_state=666)
[8]
et_clf.oob_score_
0.892
集成学习解决回归问题
[9]
from sklearn.ensemble import BaggingRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.ensemble import ExtraTreesRegressor
13-6 Ada Boosting和Gradient Boosting
Notbook 示例
Notbook 源码
Boosting
[1]
import numpy as np
import matplotlib.pyplot as plt
[2]
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500,noise=0.3, random_state=42)
[3]
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
[4]
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y,
random_state=666)
Adabosting
[5]
from sklearn.ensemble import AdaBoostClassifier
from sklearn.tree import DecisionTreeClassifier
ada_clf = AdaBoostClassifier(DecisionTreeClassifier(max_depth=2),
n_estimators=500)
ada_clf.fit(X_train, y_train)
AdaBoostClassifier(base_estimator=DecisionTreeClassifier(max_depth=2),
n_estimators=500)
[6]
ada_clf.score(X_test, y_test)
0.824
Gradient Boosting
[7]
from sklearn.ensemble import GradientBoostingClassifier
gd_clf = GradientBoostingClassifier(max_depth=2,n_estimators=30)
gd_clf.fit(X_train, y_train)
GradientBoostingClassifier(max_depth=2, n_estimators=30)
[8]
gd_clf.score(X_test,y_test)
0.848
Boosting 解决回归问题
[9]
from sklearn.ensemble import AdaBoostRegressor
from sklearn.ensemble import GradientBoostingRegressor
13-7 Stacking
第14章 更多机器学习算法
14-1 scikit-learan 文档的学习
网址：www.scikit-learn.org
