综述
“余既滋兰之九畹兮,又树蕙之百亩”
本文采用编译器:jupyter
集成学习 (ensemble learning)通过构建并结合多个学习器来完成学习任务。下图显示出集成学习的一般结构,先产生一组"个体学习器",然后再通过某种方式将他们结合起来解决一个问题。
在日常生活中也充满着集成学习的例子:比如“买东西找人推荐”“专家会诊确定病情”等 。
个体学习器由常见的机器学习模型构成,采用投票方式决定最终结果,为了达到“集成”的目的,个体学习器应具有差异性。
01 集成学习
import numpy as np
import matplotlib.pyplot as plt
from sklearn import datasets
X, y = datasets.make_moons(n_samples=500, noise=0.3, random_state=42)
plt.scatter(X[y==0,0], X[y==0,1])
plt.scatter(X[y==1,0], X[y==1,1])
plt.show()
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
# 使用逻辑回归分类
from sklearn.linear_model import LogisticRegression
log_clf = LogisticRegression()
log_clf.fit(X_train, y_train)
log_clf.score(X_test, y_test)
# Out[5]:
# 0.86399999999999999
# 使用SVM方式分类
from sklearn.svm import SVC
svm_clf = SVC()
svm_clf.fit(X_train, y_train)
svm_clf.score(X_test, y_test)
# Out[6]:
# 0.88800000000000001
# 使用决策树进行预测
from sklearn.tree import DecisionTreeClassifier
dt_clf = DecisionTreeClassifier()
dt_clf.fit(X_train, y_train)
dt_clf.score(X_test, y_test)
# Out[7]:
# 0.872
# 集成学习,少数服从多数过程
y_predict1 = log_clf.predict(X_test)
y_predict2 = svm_clf.predict(X_test)
y_predict3 = dt_clf.predict(X_test)
y_predict = np.array((y_predict1 + y_predict2 + y_predict3) >= 2, dtype='int')
y_predict[:10]
# Out[10]:
# array([1, 0, 0, 1, 1, 1, 0, 0, 0, 0])
from sklearn.metrics import accuracy_score
accuracy_score(y_test, y_predict)
# 可以看出分类准确率比三种单独的方法要高
# Out[11]:
# 0.90400000000000003
使用Voting Classifier
from sklearn.ensemble import VotingClassifier
# 创建对象时,传入估计器
voting_clf = VotingClassifier(estimators=[
('log_clf', LogisticRegression()),
('svm_clf', SVC()),
('dt_clf', DecisionTreeClassifier())
], voting='hard')
voting_clf.fit(X_train, y_train)
"""
Out[13]:
VotingClassifier(estimators=[('log_clf', LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, max_iter=100, multi_class='ovr', n_jobs=1,
penalty='l2', random_state=None, solver='liblinear', tol=0.0001,
verbose=0, warm_start=False)), ('svm_clf... min_weight_fraction_leaf=0.0, presort=False, random_state=None,
splitter='best'))],
flatten_transform=None, n_jobs=1, voting='hard', weights=None)
"""
voting_clf.score(X_test, y_test)
# Out[14]:
# 0.90400000000000003
02 Soft Voting
由于不同的学习机器确定结果所对应的概率不同,