机器学习之集成学习

集成学习

bagging

bagging:

# 导入算法包以及数据集
from sklearn import neighbors
from sklearn import datasets
from sklearn.ensemble import BaggingClassifier
from sklearn import tree
from sklearn.model_selection import train_test_split
import numpy as np
import matplotlib.pyplot as plt

# 绘制分类等高线图函数
def plot(model):
    # 获取数据值所在的范围
    x_min, x_max = x_data[:, 0].min() - 1, x_data[:, 0].max() + 1
    y_min, y_max = x_data[:, 1].min() - 1, x_data[:, 1].max() + 1
    # 生成网格矩阵
    xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),
                         np.arange(y_min, y_max, 0.02))
    z = model.predict(np.c_[xx.ravel(), yy.ravel()])
    z = z.reshape(xx.shape)
    # 等高线图
    plt.contourf(xx, yy, z)


iris = datasets.load_iris()
x_data = iris.data[:,:2] #只取前两列作为训练特征
y_data = iris.target
plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
x_train,x_test,y_train,y_test = train_test_split(x_data, y_data)

#普通Knn分类器
knn = neighbors.KNeighborsClassifier() #knn分类器
knn.fit(x_train, y_train)
# 分区
plot(knn)
# 样本散点图
plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
plt.show()
# 准确率
knn.score(x_test, y_test)

#bagging分类器结合knn，n_estimators=100表示做100次有放回的抽样
bagging_knn = BaggingClassifier(knn, n_estimators=100)#bagging分类器
# 输入数据建立模型
bagging_knn.fit(x_train, y_train)
plot(bagging_knn)
# 样本散点图
plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
plt.show()
bagging_knn.score(x_test, y_test)

普通KNN分类器：

Bagging+Knn分类器：

随机森林RF:

from sklearn import tree
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
import numpy as np
import matplotlib.pyplot as plt

#定义分类画图函数
def plot(model):
    # 获取数据值所在的范围
    x_min, x_max = x_data[:, 0].min() - 1, x_data[:, 0].max() + 1
    y_min, y_max = x_data[:, 1].min() - 1, x_data[:, 1].max() + 1
    # 生成网格矩阵
    xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),
                         np.arange(y_min, y_max, 0.02))

    z = model.predict(np.c_[xx.ravel(), yy.ravel()])# ravel与flatten类似，多维数据转一维。flatten不会改变原始数据，ravel会改变原始数据
    z = z.reshape(xx.shape)
    # 等高线图
    plt.contourf(xx, yy, z)
    # 样本散点图
    plt.scatter(x_test[:, 0], x_test[:, 1], c=y_test)
    plt.show()

# 载入数据
data = np.genfromtxt("LR-testSet2.txt", delimiter=",")
x_data = data[:,:-1]
y_data = data[:,-1]

x_train,x_test,y_train,y_test = train_test_split(x_data, y_data, test_size = 0.5)

#单棵决策树
dtree = tree.DecisionTreeClassifier()
dtree.fit(x_train, y_train)
plot(dtree)
dtree.score(x_test, y_test)

#随机森林，n_estimators=50：表示做50次bagging，建立50棵决策树。每棵决策树对树种的每一个样本都有分类，最终投票决定属于哪一类
RF = RandomForestClassifier(n_estimators=50)
RF.fit(x_train, y_train)
plot(RF)
RF.score(x_test, y_test)

单棵决策树结果：

随机森林结果：

boosting:

import numpy as np
import matplotlib.pyplot as plt
from sklearn import tree
from sklearn.ensemble import AdaBoostClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.datasets import make_gaussian_quantiles
from sklearn.metrics import classification_report

# 生成2维正态分布，生成的数据按分位数分为两类，500个样本,2个样本特征
x1, y1 = make_gaussian_quantiles(n_samples=500, n_features=2,n_classes=2)
# 生成2维正态分布，生成的数据按分位数分为两类，500个样本,2个样本特征均值都为3
x2, y2 = make_gaussian_quantiles(mean=(3, 3), n_samples=500, n_features=2, n_classes=2)
# 将两组数据合成一组数据
x_data = np.concatenate((x1, x2))
y_data = np.concatenate((y1, -y2+1)) # -y2+1表示将第二组样本的标签翻转

plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
plt.show()

# 决策树模型
model = tree.DecisionTreeClassifier(max_depth=3)
model.fit(x_data, y_data)
# 分块
x_min, x_max = x_data[:, 0].min() - 1, x_data[:, 0].max() + 1
y_min, y_max = x_data[:, 1].min() - 1, x_data[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),np.arange(y_min, y_max, 0.02))
z = model.predict(np.c_[xx.ravel(), yy.ravel()])
z = z.reshape(xx.shape)
plt.contourf(xx, yy, z)
plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
plt.show()


# AdaBoost模型，n_estimators=10：表示基分类器提升（循环）次数
model = AdaBoostClassifier(DecisionTreeClassifier(max_depth=3),n_estimators=10)
model.fit(x_data, y_data)
# 分块
x_min, x_max = x_data[:, 0].min() - 1, x_data[:, 0].max() + 1
y_min, y_max = x_data[:, 1].min() - 1, x_data[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),np.arange(y_min, y_max, 0.02))
z = model.predict(np.c_[xx.ravel(), yy.ravel()])
z = z.reshape(xx.shape)
plt.contourf(xx, yy, z)
plt.scatter(x_data[:, 0], x_data[:, 1], c=y_data)
plt.show()
# 模型准确率
model.score(x_data,y_data)

原始数据：

决策树模型：

AdaBoost模型：

stacking：

from sklearn import datasets  
from sklearn import model_selection  
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier  
from sklearn.tree import DecisionTreeClassifier
from mlxtend.classifier import StackingClassifier
from sklearn.ensemble import VotingClassifier
import numpy as np

# 载入数据集
iris = datasets.load_iris()  
# 只要第1,2列的特征
x_data, y_data = iris.data[:, 1:3], iris.target  

# 定义三个不同的分类器
clf1 = KNeighborsClassifier(n_neighbors=1)  
clf2 = DecisionTreeClassifier() 
clf3 = LogisticRegression()  
 
# 定义一个次级分类器
lr = LogisticRegression()  
print("stacking:")
sclf1 = StackingClassifier(classifiers=[clf1, clf2, clf3],meta_classifier=lr)
for clf,label in zip([clf1, clf2, clf3, sclf1],['KNN','Decision Tree','LogisticRegression','StackingClassifier']): 
    #cv=3表示将数据分成三份的交叉验证，会得到三个结果，下面求平均值
    scores = model_selection.cross_val_score(clf, x_data, y_data, cv=3, scoring='accuracy')  
    print("Accuracy: %0.2f [%s]" % (scores.mean(), label)) 

print("voting:")
sclf2 = VotingClassifier([('knn',clf1),('dtree',clf2), ('lr',clf3)])   
for clf, label in zip([clf1, clf2, clf3, sclf2],
                      ['KNN','Decision Tree','LogisticRegression','VotingClassifier']):  
  
    scores = model_selection.cross_val_score(clf, x_data, y_data, cv=3, scoring='accuracy')  
    print("Accuracy: %0.2f [%s]" % (scores.mean(), label))