第十三章：集成学习和随机森林

多种机器学习算法都能做同样的事情。让不同的算法针对同一个数据都跑一遍，最终使用投票的方法，少数服从多数，用多数投票的结果作为最终的结果。

import numpy as np
import matplotlib.pyplot as plt
from sklearn import datasets

X, y = datasets.make_moons(noise=0.25, random_state=666)

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, 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()

自己实现集成学习

逻辑回归

from sklearn.linear_model import LogisticRegression

log_clf = LogisticRegression()
log_clf.fit(X_train, y_train)
log_clf.score(X_test, y_test)

输出：0.864

SVM

from sklearn.svm import SVC

svm_clf = SVC()
svm_clf.fit(X_train, y_train)
svm_clf.score(X_test, y_test)

输出：0.888

决策树

from sklearn.tree import DecisionTreeClassifier

dt_clf = DecisionTreeClassifier()
dt_clf.fit(X_train, y_train)
dt_clf.score(X_test, y_test)

输出：0.84

集成学习

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')

from sklearn.metrics import accuracy_score

accuracy_score(y_test, y_predict)

输出：0.896 使用集成学习方法提高了准确率

使用Voting Classifier

from sklearn.ensemble import VotingClassifier

voting_clf = VotingClassifier(estimators=[
    ('log_clf', LogisticRegression()),
    ('svm_clf', SVC()),
    ('dt_clf', DecisionTreeClassifier())
], voting='hard')

voting_clf.fit(X_train, y_train)
voting_clf.score(X_test, y_test)

输出：0.896