数据挖掘之决策树
今天想分享的是数据挖掘中决策树实例,具体的原理我就不分享了,代码案在我的github上:决策树
1 决策树代码案例
该案例主要是用sklearn构建决策树的案例,选取前两个特征构建模型,并用matplotlib做模型可视化展示
#!/usr/bin/python
# -*- coding:utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
from sklearn import tree
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
def iris_type(s):
it = {b'Iris-setosa': 0, b'Iris-versicolor': 1, b'Iris-virginica': 2}
return it[s]
# 花萼长度、花萼宽度,花瓣长度,花瓣宽度
# iris_feature = 'sepal length', 'sepal width', 'petal length', 'petal width'
iris_feature = u'花萼长度', u'花萼宽度', u'花瓣长度', u'花瓣宽度'
if __name__ == "__main__":
mpl.rcParams['font.sans-serif'] = [u'SimHei']
mpl.rcParams['axes.unicode_minus'] = False
path = './iris.data' # 数据文件路径
data = np.loadtxt(path, dtype=float, delimiter=',', converters={4: iris_type})
x, y = np.split(data, (4,), axis=1)
# 为了可视化,仅使用前两列特征
x = x[:, :2]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.3, random_state=1)
#ss = StandardScaler()
#ss = ss.fit(x_train)
# 决策树参数估计
# min_samples_split = 10:如果该结点包含的样本数目大于10,则(有可能)对其分支
# min_samples_leaf = 10:若将某结点分支后,得到的每个子结点样本数目都大于10,则完成分支;否则,不进行分支
#构建模型框架
# ss = StandardScaler()
# x_train= ss.fit_transform(x_train)
# x_test= ss.fit_transform(x_test)
# model = DecisionTreeClassifier(criterion="entropy",max_depth=3)
# model = model.fit(x_train, y_train)
# y_test_hat = model.predict(x_test)
#或用管道构建模型框架
model = Pipeline([
('ss', StandardScaler()), #数据标准化过程
('DTC', DecisionTreeClassifier(criterion='entropy', max_depth=5))])
# clf = DecisionTreeClassifier(criterion='entropy', max_depth=3)
model = model.fit(x_train, y_train)
y_test_hat = model.predict(x_test) # 测试数据
# 保存
# dot -Tpng -o 1.png 1.dot
# f = open('./iris_tree.dot', 'w')
# tree.export_graphviz(model.get_params('DTC')['DTC'], out_file=f)
# f.close()
# 画图
N, M = 100, 100 # 横纵各采样多少个值
x1_min, x1_max = x[:, 0].min(), x[:, 0].max() # 第0列的范围
x2_min, x2_max = x[:, 1].min(), x[:, 1].max() # 第1列的范围
t1 = np.linspace(x1_min, x1_max, N)
t2 = np.linspace(x2_min, x2_max, M)
x1, x2 = np.meshgrid(t1, t2) # 生成网格采样点
x_show = np.stack((x1.flat, x2.flat), axis=1) # 测试点
cm_light = mpl.colors.ListedColormap(['#A0FFA0', '#FFA0A0', '#A0A0FF'])
cm_dark = mpl.colors.ListedColormap(['g', 'r', 'b'])
# x_show = ss.fit_transform(x_show)
y_show_hat = model.predict(x_show) # 预测值
y_show_hat = y_show_hat.reshape(x1.shape) # 使之与输入的形状相同
plt.figure(facecolor='w')
plt.pcolormesh(x1, x2, y_show_hat, cmap=cm_light) # 预测值的显示
plt.scatter(x_test[:, 0], x_test[:, 1], c=y_test.ravel(), edgecolors='k', s=100, cmap=cm_dark, marker='*') # 测试数据
plt.scatter(x[:, 0], x[:, 1], c=y.ravel(), edgecolors='k', s=40, cmap=cm_dark) # 全部数据
plt.xlabel(iris_feature[0], fontsize=15)
plt.ylabel(iris_feature[1], fontsize=15)
plt.xlim(x1_min, x1_max)
plt.ylim(x2_min, x2_max)
plt.grid(True)
plt.title(u'鸢尾花数据的决策树分类', fontsize=17)
plt.show()
# 训练集上的预测结果
y_test = y_test.reshape(-1)
print (y_test_hat)
print (y_test)
result = (y_test_hat == y_test) # True则预测正确,False则预测错误
acc = np.mean(result)
print ('准确度: %.2f%%' % (100 * acc))
接下来我们再画一个随着决策树深度(超参数)变化,在测试集上错误率变化情况
# 过拟合:错误率
depth = np.arange(1, 15)
err_list = []
for d in depth:
clf = DecisionTreeClassifier(criterion='entropy', max_depth=d)
clf = clf.fit(x_train, y_train)
y_test_hat = clf.predict(x_test) # 测试数据
result = (y_test_hat == y_test) # True则预测正确,False则预测错误
err = 1 - np.mean(result)
err_list.append(err)
print (d, ' 错误率: %.2f%%' % (100 * err))
plt.figure(facecolor='w')
plt.plot(depth, err_list, 'ro-', lw=2)
plt.xlabel(u'决策树深度', fontsize=15)
plt.ylabel(u'错误率', fontsize=15)
plt.title(u'决策树深度与过拟合', fontsize=17)
plt.grid(True)
plt.show()
2 选择不同的特征构建决策树
#!/usr/bin/python
# -*- coding:utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
from sklearn.tree import DecisionTreeClassifier
def iris_type(s):
it = {b'Iris-setosa': 0, b'Iris-versicolor': 1, b'Iris-virginica': 2}
return it[s]
# 'sepal length', 'sepal width', 'petal length', 'petal width'
iris_feature = u'花萼长度', u'花萼宽度', u'花瓣长度', u'花瓣宽度'
if __name__ == "__main__":
mpl.rcParams['font.sans-serif'] = [u'SimHei'] # 黑体 FangSong/KaiTi
mpl.rcParams['axes.unicode_minus'] = False
path = './iris.data' # 数据文件路径
data = np.loadtxt(path, dtype=float, delimiter=',', converters={4: iris_type})
x_prime, y = np.split(data, (4,), axis=1)
feature_pairs = [[0, 1], [0, 2], [0, 3], [1, 2], [1, 3], [2, 3]]
plt.figure(figsize=(10, 9), facecolor='#FFFFFF')
for i, pair in enumerate(feature_pairs):
# 准备数据
x = x_prime[:, pair]
# 决策树学习
clf = DecisionTreeClassifier(criterion='entropy', min_samples_leaf=3)
dt_clf = clf.fit(x, y)
# 画图
N, M = 500, 500 # 横纵各采样多少个值
x1_min, x1_max = x[:, 0].min(), x[:, 0].max() # 第0列的范围
x2_min, x2_max = x[:, 1].min(), x[:, 1].max() # 第1列的范围
t1 = np.linspace(x1_min, x1_max, N)
t2 = np.linspace(x2_min, x2_max, M)
x1, x2 = np.meshgrid(t1, t2) # 生成网格采样点
x_test = np.stack((x1.flat, x2.flat), axis=1) # 测试点
# 训练集上的预测结果
y_hat = dt_clf.predict(x)
y = y.reshape(-1)
c = np.count_nonzero(y_hat == y) # 统计预测正确的个数
print ('特征: ', iris_feature[pair[0]], ' + ', iris_feature[pair[1]],)
print ('预测正确数目:', c,)
print ('准确率: %.2f%%' % (100 * float(c) / float(len(y))))
# 显示
cm_light = mpl.colors.ListedColormap(['#A0FFA0', '#FFA0A0', '#A0A0FF'])
cm_dark = mpl.colors.ListedColormap(['g', 'r', 'b'])
y_hat = dt_clf.predict(x_test) # 预测值
y_hat = y_hat.reshape(x1.shape) # 使之与输入的形状相同
plt.subplot(2, 3, i+1)
plt.pcolormesh(x1, x2, y_hat, cmap=cm_light) # 预测值
plt.scatter(x[:, 0], x[:, 1], c=y, edgecolors='k', cmap=cm_dark) # 样本
plt.xlabel(iris_feature[pair[0]], fontsize=14)
plt.ylabel(iris_feature[pair[1]], fontsize=14)
plt.xlim(x1_min, x1_max)
plt.ylim(x2_min, x2_max)
plt.grid()
plt.suptitle(u'决策树对鸢尾花数据的两特征组合的分类结果', fontsize=18)
plt.tight_layout(2) #图像外部边缘的调整
plt.subplots_adjust(top=0.92) #图像间的空白区域
plt.show()
3 决策树回归
我们也可以利用决策树做连续值预测的回归
!/usr/bin/python
# -*- coding:utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
from sklearn.tree import DecisionTreeRegressor
if __name__ == "__main__":
N = 100
x = np.random.rand(N) * 6 - 3 # [-3,3)
x.sort()
y = np.sin(x) + np.random.randn(N) * 0.05
print (y)
x = x.reshape(-1, 1) # 转置后,得到N个样本,每个样本都是1维的
print (x)
reg = DecisionTreeRegressor(criterion='mse', max_depth=9) #mse使用均方误差作为分类准则
dt = reg.fit(x, y)
x_test = np.linspace(-3, 3, 50).reshape(-1, 1)
y_hat = dt.predict(x_test)
plt.plot(x, y, 'r*', linewidth=2, label='Actual')
plt.plot(x_test, y_hat, 'g-', linewidth=2, label='Predict')
plt.legend(loc='upper left')
plt.grid()
plt.show()
# 比较决策树的深度影响
depth = [2, 4, 6, 8, 10]
clr = 'rgbmy'
reg = [DecisionTreeRegressor(criterion='mse', max_depth=depth[0]),
DecisionTreeRegressor(criterion='mse', max_depth=depth[1]),
DecisionTreeRegressor(criterion='mse', max_depth=depth[2]),
DecisionTreeRegressor(criterion='mse', max_depth=depth[3]),
DecisionTreeRegressor(criterion='mse', max_depth=depth[4])]
plt.plot(x, y, 'k^', linewidth=2, label='Actual')
x_test = np.linspace(-3, 3, 50).reshape(-1, 1)
for i, r in enumerate(reg):
dt = r.fit(x, y)
y_hat = dt.predict(x_test)
plt.plot(x_test, y_hat, '-', color=clr[i], linewidth=2, label='Depth=%d' % depth[i])
plt.legend(loc='upper left')
plt.grid()
plt.show()
4 多label的决策树回归
有时候我们要更具一个样本预测多个label值的方案,因此,我们可以构建多label的决策树回归。
import numpy as np
import matplotlib.pyplot as plt
from sklearn.tree import DecisionTreeRegressor
if __name__ == "__main__":
N = 300
x = np.random.rand(N) * 8 - 4 # [-4,4)
x.sort()
y1 = 16*np.sin(x)** 3 + np.random.randn(N)*0.1
y2 = 13*np.cos(x) -5*np.cos(2*x)-2*np.cos(3*x)-np.cos(4*x)+ np.random.randn(N) * 0.01
# y1 = np.sin(x) + np.random.randn(N) * 0.05
# y2 = np.cos(x) + np.random.randn(N) * 0.1
y = np.vstack((y1, y2))
y = np.vstack((y1, y2)).T
x = x.reshape(-1, 1) # 转置后,得到N个样本,每个样本都是1维的
deep = 3
reg = DecisionTreeRegressor(criterion='mse', max_depth=deep)
dt = reg.fit(x, y)
x_test = np.linspace(-4, 4, num=1000).reshape(-1, 1)
print (x_test)
y_hat = dt.predict(x_test)
print (y_hat)
plt.scatter(y[:, 0], y[:, 1], c='r', s=40, label='Actual')
plt.scatter(y_hat[:, 0], y_hat[:, 1], c='g', marker='s', s=100, label='Depth=%d' % deep, alpha=1)
plt.legend(loc='upper left')
plt.xlabel('y1')
plt.ylabel('y2')
plt.grid()
plt.show()
随机森林
随机森林是一种集成的学习方法,它与另一种集成学习bagging算法区别在于:
bagging算法:
- 从原始样本集中抽取训练集。每轮从原始样本集中使用Bootstraping的方法抽取n个训练样本(在训练集中,有些样本可能被多次抽取到,而有些样本可能一次都没有被抽中)。共进行k轮抽取,得到k个训练集。(k个训练集之间是相互独立的)
- 每次使用一个训练集得到一个模型,k个训练集共得到k个模型。(注:这里并没有具体的分类算法或回归方法,我们可以根据具体问题采用不同的分类或回归方法,如决策树、感知器等)
- 对分类问题:将上步得到的k个模型采用投票的方式得到分类结果;对回归问题,计算上述模型的均值作为最后的结果。(所有模型的重要性相同)
RandomFroest算法:
- 假设我们设定训练集中的样本个数为N ,然后通过 Bootstrap Sampling 来获得 N 个有重复的样本集,这样的抽样结果将作为我们生成决策树的训练集;
- 对于有 d 个特征的数据集,每个节点都将随机选择 k (k<d) 个特定的变量,然后运用这 k 个变量来确定最佳的分裂点。在决策树的生成过程中,k 的值是保持不变的, 随机选取特征会增加树的独立性;
- 每棵决策树都最大可能地进行生长而不进行剪枝;
- 通过对所有的决策树进行加总来预测新的数据(在分类时采用多数投票,在回归时采用平均)。
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
from sklearn.ensemble import RandomForestClassifier
def iris_type(s):
it = {b'Iris-setosa': 0, b'Iris-versicolor': 1, b'Iris-virginica': 2}
return it[s]
# 'sepal length', 'sepal width', 'petal length', 'petal width'
iris_feature = u'花萼长度', u'花萼宽度', u'花瓣长度', u'花瓣宽度'
if __name__ == "__main__":
mpl.rcParams['font.sans-serif'] = [u'SimHei'] # 黑体 FangSong/KaiTi
mpl.rcParams['axes.unicode_minus'] = False
path = './iris.data' # 数据文件路径
data = np.loadtxt(path, dtype=float, delimiter=',', converters={4: iris_type})
x_prime, y = np.split(data, (4,), axis=1)
feature_pairs = [[0, 1], [0, 2], [0, 3], [1, 2], [1, 3], [2, 3]]
plt.figure(figsize=(10, 9), facecolor='#FFFFFF')
for i, pair in enumerate(feature_pairs):
# 准备数据
x = x_prime[:, pair]
# 随机森林
clf = RandomForestClassifier(n_estimators=50, criterion='entropy', max_depth=3)
rf_clf = clf.fit(x, y.ravel())
# 画图
N, M = 100, 100 # 横纵各采样多少个值
x1_min, x1_max = x[:, 0].min(), x[:, 0].max() # 第0列的范围
x2_min, x2_max = x[:, 1].min(), x[:, 1].max() # 第1列的范围
t1 = np.linspace(x1_min, x1_max, N)
t2 = np.linspace(x2_min, x2_max, M)
x1, x2 = np.meshgrid(t1, t2) # 生成网格采样点
x_test = np.stack((x1.flat, x2.flat), axis=1) # 测试点
# 训练集上的预测结果
y_hat = rf_clf.predict(x)
y = y.reshape(-1)
c = np.count_nonzero(y_hat == y) # 统计预测正确的个数
print ('特征: ', iris_feature[pair[0]], ' + ', iris_feature[pair[1]],)
print ('\t预测正确数目:', c,)
print ('\t准确率: %.2f%%' % (100 * float(c) / float(len(y))))
# 显示
cm_light = mpl.colors.ListedColormap(['#A0FFA0', '#FFA0A0', '#A0A0FF'])
cm_dark = mpl.colors.ListedColormap(['g', 'r', 'b'])
y_hat = rf_clf.predict(x_test) # 预测值
y_hat = y_hat.reshape(x1.shape) # 使之与输入的形状相同
plt.subplot(2, 3, i+1)
plt.pcolormesh(x1, x2, y_hat, cmap=cm_light) # 预测值
plt.scatter(x[:, 0], x[:, 1], c=y, edgecolors='k', cmap=cm_dark) # 样本
plt.xlabel(iris_feature[pair[0]], fontsize=14)
plt.ylabel(iris_feature[pair[1]], fontsize=14)
plt.xlim(x1_min, x1_max)
plt.ylim(x2_min, x2_max)
plt.grid()
plt.tight_layout(2.5)
plt.subplots_adjust(top=0.92)
plt.suptitle(u'随机森林对鸢尾花数据的两特征组合的分类结果', fontsize=18)
plt.show()
6 利用GridSearchCV对超参数优化
我们也可以用机器学习sklearn利用GridSearchCV进行超参数优化,我们建立一个新的脚本文件
#!/usr/bin/python
# -*- coding:utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
from sklearn import tree
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.model_selection import ShuffleSplit
from sklearn.metrics import classification_report
def iris_type(s):
it = {b'Iris-setosa': 0, b'Iris-versicolor': 1, b'Iris-virginica': 2}
return it[s]
# 花萼长度、花萼宽度,花瓣长度,花瓣宽度
# iris_feature = 'sepal length', 'sepal width', 'petal length', 'petal width'
iris_feature = u'花萼长度', u'花萼宽度', u'花瓣长度', u'花瓣宽度'
if __name__ == "__main__":
mpl.rcParams['font.sans-serif'] = [u'SimHei']
mpl.rcParams['axes.unicode_minus'] = False
path = './iris.data' # 数据文件路径
data = np.loadtxt(path, dtype=float, delimiter=',', converters={4: iris_type})
x, y = np.split(data, (4,), axis=1)
# 为了可视化,仅使用前两列特征
x = x[:, :2]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.3, random_state=1)
# ss = StandardScaler()
# ss = ss.fit(x_train)
# 或用管道构建模型框架
# 测试数据
# 最优超参数组合列表
# cv_split = ShuffleSplit(n_splits=5, train_size=0.7, test_size=0.25)
pip_count = Pipeline(
[('ss', StandardScaler()), # 数据标准化过程
('DTC', DecisionTreeClassifier())])
params = {'DTC__max_depth': [2, 3, 4, 5],
'DTC__criterion': ['gini', 'entropy']
}
model = GridSearchCV(pip_count ,
params,
refit=True,
return_train_score=True, # 后续版本需要指定True才有score方法
cv=4)
model = model.fit(x_train, y_train)
print('best_estimator_:\n', model.best_estimator_) #查看表格搜索最优参数
print('best_socre_:\n',model.best_score_)
print('best_params_:\n', model.best_params_)
best_model = model.best_estimator_
predict_y = best_model.predict(x_test)
print("classification_report:\n",classification_report(y_test, predict_y))
print("best_score:\n",best_model.score(x_test,y_test))