机器学习之决策树三-CART原理与代码实现

可紊 提交于 2021-02-18 17:53:04

                                             决策树系列三—CART原理与代码实现

                                                  本文系作者原创,转载请注明出处:https://www.cnblogs.com/further-further-further/p/9482885.html 

ID3,C4.5算法缺点

ID3决策树可以有多个分支,但是不能处理特征值为连续的情况。

在ID3中,每次根据“最大信息熵增益”选取当前最佳的特征来分割数据,并按照该特征的所有取值来切分,

也就是说如果一个特征有4种取值,数据将被切分4份,一旦按某特征切分后,该特征在之后的算法执行中,

将不再起作用,所以有观点认为这种切分方式过于迅速。

C4.5中是用信息增益比率(gain ratio)来作为选择分支的准则。和ID3一样,C4.5算法分类结果存在过拟合。

为了解决过拟合问题,这里介绍一种新的算法CART。

CART(classification and regression tree)

CART由特征选择、树的生成及剪枝组成,既可以用于分类也可以用于回归。

分类:如晴天/阴天/雨天、用户性别、邮件是否是垃圾邮件; 

回归:预测实数值,如明天的温度、用户的年龄等; 

 

CART决策树的生成就是递归地构建二叉决策树的过程,对分类、以及剪枝采用信息增益最大化准则,这里信息增益采用的基尼指数公式,

当然也可以使用ID3的信息熵公式算法。

基尼指数

分类问题中,假设有K个类别,样本点属于第k类的概率为p_k,则概率分布的基尼指数定义为

                  

 

对于给定的样本集合D,其基尼指数为

                  

 

生成的二叉树类似于

      

剪枝算法

CART剪枝算法从“完全生长”的决策树的底端减去一些子树,是决策树变小(模型变简单),从而能够对未知数据有更准确的预测,防止过拟合。

后剪枝需要从训练集生成一棵完整的决策树,然后自底向上对非叶子节点进行考察。利用信息增益与给定阈值判断是否将该节点对应的子树替换成叶节点。

   

 

代码实现

每个函数算法我基本上都做了较为详细的注释,希望对大家理解算法原理有所帮助。

因为没有上传附件功能,只能用笨办法。将原始数据复制到本地txt文件中,然后将txt格式改成dataSet.csv文件,

放在代码文件所在的路径。

  1 SepalLength,SepalWidth,PetalLength,PetalWidth,Name
  2 5.1,3.5,1.4,0.2,setosa
  3 4.9,3,1.4,0.2,setosa
  4 4.7,3.2,1.3,0.2,setosa
  5 4.6,3.1,1.5,0.2,setosa
  6 5,3.6,1.4,0.2,setosa
  7 5.4,3.9,1.7,0.4,setosa
  8 4.6,3.4,1.4,0.3,setosa
  9 5,3.4,1.5,0.2,setosa
 10 4.4,2.9,1.4,0.2,setosa
 11 4.9,3.1,1.5,0.1,setosa
 12 5.4,3.7,1.5,0.2,setosa
 13 4.8,3.4,1.6,0.2,setosa
 14 4.8,3,1.4,0.1,setosa
 15 4.3,3,1.1,0.1,setosa
 16 5.8,4,1.2,0.2,setosa
 17 5.7,4.4,1.5,0.4,setosa
 18 5.4,3.9,1.3,0.4,setosa
 19 5.1,3.5,1.4,0.3,setosa
 20 5.7,3.8,1.7,0.3,setosa
 21 5.1,3.8,1.5,0.3,setosa
 22 5.4,3.4,1.7,0.2,setosa
 23 5.1,3.7,1.5,0.4,setosa
 24 4.6,3.6,1,0.2,setosa
 25 5.1,3.3,1.7,0.5,setosa
 26 4.8,3.4,1.9,0.2,setosa
 27 5,3,1.6,0.2,setosa
 28 5,3.4,1.6,0.4,setosa
 29 5.2,3.5,1.5,0.2,setosa
 30 5.2,3.4,1.4,0.2,setosa
 31 4.7,3.2,1.6,0.2,setosa
 32 4.8,3.1,1.6,0.2,setosa
 33 5.4,3.4,1.5,0.4,setosa
 34 5.2,4.1,1.5,0.1,setosa
 35 5.5,4.2,1.4,0.2,setosa
 36 4.9,3.1,1.5,0.1,setosa
 37 5,3.2,1.2,0.2,setosa
 38 5.5,3.5,1.3,0.2,setosa
 39 4.9,3.1,1.5,0.1,setosa
 40 4.4,3,1.3,0.2,setosa
 41 5.1,3.4,1.5,0.2,setosa
 42 5,3.5,1.3,0.3,setosa
 43 4.5,2.3,1.3,0.3,setosa
 44 4.4,3.2,1.3,0.2,setosa
 45 5,3.5,1.6,0.6,setosa
 46 5.1,3.8,1.9,0.4,setosa
 47 4.8,3,1.4,0.3,setosa
 48 5.1,3.8,1.6,0.2,setosa
 49 4.6,3.2,1.4,0.2,setosa
 50 5.3,3.7,1.5,0.2,setosa
 51 5,3.3,1.4,0.2,setosa
 52 7,3.2,4.7,1.4,versicolor
 53 6.4,3.2,4.5,1.5,versicolor
 54 6.9,3.1,4.9,1.5,versicolor
 55 5.5,2.3,4,1.3,versicolor
 56 6.5,2.8,4.6,1.5,versicolor
 57 5.7,2.8,4.5,1.3,versicolor
 58 6.3,3.3,4.7,1.6,versicolor
 59 4.9,2.4,3.3,1,versicolor
 60 6.6,2.9,4.6,1.3,versicolor
 61 5.2,2.7,3.9,1.4,versicolor
 62 5,2,3.5,1,versicolor
 63 5.9,3,4.2,1.5,versicolor
 64 6,2.2,4,1,versicolor
 65 6.1,2.9,4.7,1.4,versicolor
 66 5.6,2.9,3.6,1.3,versicolor
 67 6.7,3.1,4.4,1.4,versicolor
 68 5.6,3,4.5,1.5,versicolor
 69 5.8,2.7,4.1,1,versicolor
 70 6.2,2.2,4.5,1.5,versicolor
 71 5.6,2.5,3.9,1.1,versicolor
 72 5.9,3.2,4.8,1.8,versicolor
 73 6.1,2.8,4,1.3,versicolor
 74 6.3,2.5,4.9,1.5,versicolor
 75 6.1,2.8,4.7,1.2,versicolor
 76 6.4,2.9,4.3,1.3,versicolor
 77 6.6,3,4.4,1.4,versicolor
 78 6.8,2.8,4.8,1.4,versicolor
 79 6.7,3,5,1.7,versicolor
 80 6,2.9,4.5,1.5,versicolor
 81 5.7,2.6,3.5,1,versicolor
 82 5.5,2.4,3.8,1.1,versicolor
 83 5.5,2.4,3.7,1,versicolor
 84 5.8,2.7,3.9,1.2,versicolor
 85 6,2.7,5.1,1.6,versicolor
 86 5.4,3,4.5,1.5,versicolor
 87 6,3.4,4.5,1.6,versicolor
 88 6.7,3.1,4.7,1.5,versicolor
 89 6.3,2.3,4.4,1.3,versicolor
 90 5.6,3,4.1,1.3,versicolor
 91 5.5,2.5,4,1.3,versicolor
 92 5.5,2.6,4.4,1.2,versicolor
 93 6.1,3,4.6,1.4,versicolor
 94 5.8,2.6,4,1.2,versicolor
 95 5,2.3,3.3,1,versicolor
 96 5.6,2.7,4.2,1.3,versicolor
 97 5.7,3,4.2,1.2,versicolor
 98 5.7,2.9,4.2,1.3,versicolor
 99 6.2,2.9,4.3,1.3,versicolor
100 5.1,2.5,3,1.1,versicolor
101 5.7,2.8,4.1,1.3,versicolor
102 6.3,3.3,6,2.5,virginica
103 5.8,2.7,5.1,1.9,virginica
104 7.1,3,5.9,2.1,virginica
105 6.3,2.9,5.6,1.8,virginica
106 6.5,3,5.8,2.2,virginica
107 7.6,3,6.6,2.1,virginica
108 4.9,2.5,4.5,1.7,virginica
109 7.3,2.9,6.3,1.8,virginica
110 6.7,2.5,5.8,1.8,virginica
111 7.2,3.6,6.1,2.5,virginica
112 6.5,3.2,5.1,2,virginica
113 6.4,2.7,5.3,1.9,virginica
114 6.8,3,5.5,2.1,virginica
115 5.7,2.5,5,2,virginica
116 5.8,2.8,5.1,2.4,virginica
117 6.4,3.2,5.3,2.3,virginica
118 6.5,3,5.5,1.8,virginica
119 7.7,3.8,6.7,2.2,virginica
120 7.7,2.6,6.9,2.3,virginica
121 6,2.2,5,1.5,virginica
122 6.9,3.2,5.7,2.3,virginica
123 5.6,2.8,4.9,2,virginica
124 7.7,2.8,6.7,2,virginica
125 6.3,2.7,4.9,1.8,virginica
126 6.7,3.3,5.7,2.1,virginica
127 7.2,3.2,6,1.8,virginica
128 6.2,2.8,4.8,1.8,virginica
129 6.1,3,4.9,1.8,virginica
130 6.4,2.8,5.6,2.1,virginica
131 7.2,3,5.8,1.6,virginica
132 7.4,2.8,6.1,1.9,virginica
133 7.9,3.8,6.4,2,virginica
134 6.4,2.8,5.6,2.2,virginica
135 6.3,2.8,5.1,1.5,virginica
136 6.1,2.6,5.6,1.4,virginica
137 7.7,3,6.1,2.3,virginica
138 6.3,3.4,5.6,2.4,virginica
139 6.4,3.1,5.5,1.8,virginica
140 6,3,4.8,1.8,virginica
141 6.9,3.1,5.4,2.1,virginica
142 6.7,3.1,5.6,2.4,virginica
143 6.9,3.1,5.1,2.3,virginica
144 5.8,2.7,5.1,1.9,virginica
145 6.8,3.2,5.9,2.3,virginica
146 6.7,3.3,5.7,2.5,virginica
147 6.7,3,5.2,2.3,virginica
148 6.3,2.5,5,1.9,virginica
149 6.5,3,5.2,2,virginica
150 6.2,3.4,5.4,2.3,virginica
151 5.9,3,5.1,1.8,virginica
原始数据
  1 # -*- coding: utf-8 -*-
  2 """
  3 Created on Tue Aug 14 17:36:57 2018
  4 
  5 @author: weixw
  6 """
  7 import numpy as np
  8 #定义树结构,采用的二叉树,左子树:条件为true,右子树:条件为false
  9 #leftBranch:左子树结点
 10 #rightBranch:右子树结点
 11 #col:信息增益最大时对应的列索引
 12 #value:最优列索引下,划分数据类型的值
 13 #results:分类结果
 14 #summary:信息增益最大时样本信息
 15 #data:信息增益最大时数据集
 16 class Tree:
 17     def __init__(self, leftBranch =None, rightBranch= None, col =-1, value =None, results =None, summary =None, data =None):
 18         self.leftBranch = leftBranch
 19         self.rightBranch = rightBranch
 20         self.col = col
 21         self.value = value
 22         self.results = results
 23         self.summary = summary
 24         self.data = data
 25         
 26     def __str__(self):
 27         print(u"列号:%d"%self.col)
 28         print(u"列划分值:%s"%self.value)
 29         print(u"样本信息:%s"%self.summary)
 30         return ""
 31 
 32         
 33 
 34 #划分数据集
 35 def splitDataSet(dataSet, value, column):
 36     leftList=[]
 37     rightList=[]
 38     #判断value是否是数值型
 39     if(isinstance(value, int) or isinstance(value, float)):
 40         #遍历每一行数据
 41         for rowData in dataSet:
 42             #如果某一行指定列值>=value,则将该行数据保存在leftList中,否则保存在rightList中
 43             if(rowData[column] >= value):
 44                 leftList.append(rowData)
 45             else:
 46                 rightList.append(rowData)
 47     #value为标称型
 48     else:
 49         #遍历每一行数据
 50         for rowData in dataSet:
 51             #如果某一行指定列值==value,则将该行数据保存在leftList中,否则保存在rightList中
 52             if(rowData[column] == value):
 53                 leftList.append(rowData)
 54             else:
 55                 rightList.append(rowData)
 56     return leftList, rightList
 57 
 58 #统计标签类每个样本个数
 59 '''
 60 该函数是计算gini值的辅助函数,假设输入的dataSet为为['A', 'B', 'C', 'A', 'A', 'D'],
 61 则输出为['A':3,' B':1, 'C':1, 'D':1],这样分类统计dataSet中每个类别的数量
 62 '''      
 63 def calculateDiffCount(dataSet):   
 64     results = {}
 65     for data in dataSet:
 66         # data[-1] 是数据集最后一列,也就是标签类
 67         if data[-1] not in results:
 68             results.setdefault(data[-1], 1)
 69         else:
 70             results[data[-1]] += 1
 71     return results
 72 
 73 
 74 #基尼指数公式实现
 75 def gini(dataSet):
 76     # 计算gini的值(Calculate GINI)
 77     #数据所有行
 78     length = len(dataSet)
 79     #标签列合并后的数据集
 80     results = calculateDiffCount(dataSet)
 81     imp = 0.0
 82     for i in results:
 83         imp += results[i] / length * results[i] / length
 84     return 1 - imp
 85 
 86 #生成决策树
 87 '''算法步骤'''
 88 '''根据训练数据集,从根结点开始,递归地对每个结点进行以下操作,构建二叉决策树:
 89 1 设结点的训练数据集为D,计算现有特征对该数据集的信息增益。此时,对每一个特征A,对其可能取的
 90   每个值a,根据样本点对A >=a 的测试为“是”或“否”将D分割成D1和D2两部分,利用基尼指数计算信息增益。
 91 2 在所有可能的特征A以及它们所有可能的切分点a中,选择信息增益最大的特征及其对应的切分点作为最优特征
 92   与最优切分点,依据最优特征与最优切分点,从现结点生成两个子结点,将训练数据集依特征分配到两个子结点中去。
 93 3 对两个子结点递归地调用1,2,直至满足停止条件。
 94 4 生成CART决策树。
 95 '''''''''''''''''''''
 96 #evaluationFunc= gini :采用的是基尼指数来衡量信息关注度          
 97 def buildDecisionTree(dataSet, evaluationFunc = gini):
 98     #计算基础数据集的基尼指数
 99     baseGain = evaluationFunc(dataSet)
100     #计算每一行的长度(也就是列总数)
101     columnLength = len(dataSet[0])
102     #计算数据项总数
103     rowLength = len(dataSet)
104     #初始化
105     bestGain = 0.0 #信息增益最大值
106     bestValue = None #信息增益最大时的列索引,以及划分数据集的样本值
107     bestSet = None # 信息增益最大,听过样本值划分数据集后的数据子集
108     #标签列除外(最后一列),遍历每一列数据
109     for col in range(columnLength -1):
110         #获取指定列数据
111         colSet = [example[col] for example in dataSet]
112         #获取指定列样本唯一值
113         uniqueColSet = set(colSet)
114         #遍历指定列样本集
115         for value in uniqueColSet: 
116             #分割数据集
117             leftDataSet, rightDataSet = splitDataSet(dataSet, value, col)
118             #计算子数据集概率,python3 "/"除号结果为小数
119             prop = len(leftDataSet)/rowLength
120             #计算信息增益
121             infoGain = baseGain - prop*evaluationFunc(leftDataSet) - (1 - prop)*evaluationFunc(rightDataSet)
122             #找出信息增益最大时的列索引,value,数据子集
123             if(infoGain > bestGain):
124                 bestGain = infoGain
125                 bestValue = (col, value)
126                 bestSet = (leftDataSet, rightDataSet)
127     #结点信息
128 #    nodeDescription = {'impurity:%.3f'%baseGain,'sample:%d'%rowLength}
129     nodeDescription = {'impurity': '%.3f' % baseGain, 'sample': '%d' % rowLength}
130     #数据行标签类别不一致,可以继续分类
131     #递归必须有终止条件
132     if bestGain > 0:
133         #递归,生成左子树结点,右子树结点
134         leftBranch = buildDecisionTree(bestSet[0], evaluationFunc)
135         rightBranch = buildDecisionTree(bestSet[1], evaluationFunc)
136         return Tree(leftBranch = leftBranch, rightBranch = rightBranch, col = bestValue[0]
137                     , value = bestValue[1], summary = nodeDescription, data = bestSet)
138     else:
139         #数据行标签类别都相同,分类终止
140         return Tree(results = calculateDiffCount(dataSet), summary = nodeDescription, data = dataSet)
141     
142 def createTree(dataSet, evaluationFunc=gini):
143     # 递归建立决策树, 当gain=0,时停止回归
144     #计算基础数据集的基尼指数
145     baseGain = evaluationFunc(dataSet)
146     #计算每一行的长度(也就是列总数)
147     columnLength = len(dataSet[0])
148     #计算数据项总数
149     rowLength = len(dataSet)
150     #初始化
151     bestGain = 0.0 #信息增益最大值
152     bestValue = None #信息增益最大时的列索引,以及划分数据集的样本值
153     bestSet = None # 信息增益最大,听过样本值划分数据集后的数据子集
154     #标签列除外(最后一列),遍历每一列数据
155     for col in range(columnLength -1):
156         #获取指定列数据
157         colSet = [example[col] for example in dataSet]
158         #获取指定列样本唯一值
159         uniqueColSet = set(colSet)
160         #遍历指定列样本集
161         for value in uniqueColSet: 
162             #分割数据集
163             leftDataSet, rightDataSet = splitDataSet(dataSet, value, col)
164             #计算子数据集概率,python3 "/"除号结果为小数
165             prop = len(leftDataSet)/rowLength
166             #计算信息增益
167             infoGain = baseGain - prop*evaluationFunc(leftDataSet) - (1 - prop)*evaluationFunc(rightDataSet)
168             #找出信息增益最大时的列索引,value,数据子集
169             if(infoGain > bestGain):
170                 bestGain = infoGain
171                 bestValue = (col, value)
172                 bestSet = (leftDataSet, rightDataSet)
173                 
174     impurity = u'%.3f' % baseGain
175     sample = '%d' % rowLength
176    
177     if bestGain > 0:                
178         bestFeatLabel =u'serial:%s\nimpurity:%s\nsample:%s'%(bestValue[0], impurity,sample) 
179         myTree = {bestFeatLabel:{}}
180         myTree[bestFeatLabel][bestValue[1]] = createTree(bestSet[0], evaluationFunc)
181         myTree[bestFeatLabel]['no'] = createTree(bestSet[1], evaluationFunc) 
182         return myTree
183     else:#递归需要返回值
184         bestFeatValue =u'%s\nimpurity:%s\nsample:%s'%(str(calculateDiffCount(dataSet)), impurity,sample)
185         return bestFeatValue
186     
187 #分类测试:
188 '''根据给定测试数据遍历二叉树,找到符合条件的叶子结点'''
189 '''例如测试数据为[5.9,3,4.2,1.75],按照训练数据生成的决策树分类的顺序为
190    第2列对应测试数据4.2 =>与决策树根结点(2)的value(3)比较,>=3则遍历左子树,否则遍历右子树,
191    叶子结点就是结果'''       
192 def classify(data, tree):
193     #判断是否是叶子结点,是就返回叶子结点相关信息,否就继续遍历
194     if tree.results != None:
195         return u"%s\n%s"%(tree.results, tree.summary)
196     else:
197         branch = None
198         v = data[tree.col]
199         #数值型数据
200         if isinstance(v, int) or isinstance(v, float):
201             if v >= tree.value:
202                 branch = tree.leftBranch
203             else:
204                 branch = tree.rightBranch
205         else:#标称型数据
206             if v == tree.value:
207                 branch = tree.leftBranch
208             else:
209                 branch = tree.rightBranch
210         return classify(data, branch) 
211     
212 def loadCSV(fileName):
213     def convertTypes(s):
214         s = s.strip()
215         try:
216             return float(s) if '.' in s else int(s)
217         except ValueError:
218             return s
219     data = np.loadtxt(fileName, dtype='str', delimiter=',')
220     data = data[1:, :]
221     dataSet =([[convertTypes(item) for item in row] for row in data])
222     return dataSet
223 
224 #多数表决器
225 #列中相同值数量最多为结果
226 def majorityCnt(classList):
227     import operator
228     classCounts = {}
229     for value in classList:
230         if(value not in classCounts.keys()):
231             classCounts[value] = 0
232         classCounts[value] +=1
233     sortedClassCount = sorted(classCounts.items(),key = operator.itemgetter(1),reverse =True)
234     return sortedClassCount[0][0]
235 
236 #剪枝算法(前序遍历方式:根=>左子树=>右子树)
237 '''算法步骤
238 1. 从二叉树的根结点出发,递归调用剪枝算法,直至左、右结点都是叶子结点
239 2. 计算父节点(子结点为叶子结点)的信息增益infoGain
240 3. 如果infoGain < miniGain,则选取样本多的叶子结点来取代父节点
241 4. 循环1,2,3,直至遍历完整棵树
242 '''''''''
243 def prune(tree, miniGain, evaluationFunc = gini):
244    print(u"当前结点信息:")
245    print(str(tree))
246    #如果当前结点的左子树不是叶子结点,遍历左子树
247    if(tree.leftBranch.results == None):
248        print(u"左子树结点信息:")
249        print(str(tree.leftBranch))
250        prune(tree.leftBranch, miniGain, evaluationFunc)
251    #如果当前结点的右子树不是叶子结点,遍历右子树
252    if(tree.rightBranch.results == None):
253        print(u"右子树结点信息:")
254        print(str(tree.rightBranch))
255        prune(tree.rightBranch, miniGain, evaluationFunc)
256    #左子树和右子树都是叶子结点
257    if(tree.leftBranch.results != None and tree.rightBranch.results != None):
258        #计算左叶子结点数据长度
259        leftLen = len(tree.leftBranch.data)
260        #计算右叶子结点数据长度
261        rightLen = len(tree.rightBranch.data)
262        #计算左叶子结点概率
263        leftProp = leftLen/(leftLen + rightLen)
264        #计算该结点的信息增益(子类是叶子结点)
265        infoGain = (evaluationFunc(tree.leftBranch.data + tree.rightBranch.data) - 
266                    leftProp*evaluationFunc(tree.leftBranch.data) - (1 - leftProp)*evaluationFunc(tree.rightBranch.data))
267        #信息增益 < 给定阈值,则说明叶子结点与其父结点特征差别不大,可以剪枝
268        if(infoGain < miniGain):
269            #合并左右叶子结点数据
270            dataSet = tree.leftBranch.data + tree.rightBranch.data
271            #获取标签列
272            classLabels = [example[-1] for example in dataSet]
273            #找到样本最多的标签值
274            keyLabel = majorityCnt(classLabels)
275            #判断标签值是左右叶子结点哪一个
276            if keyLabel in tree.leftBranch.results:
277                #左叶子结点取代父结点
278                tree.data = tree.leftBranch.data
279                tree.results = tree.leftBranch.results
280                tree.summary = tree.leftBranch.summary
281            else:
282                #右叶子结点取代父结点
283                tree.data = tree.rightBranch.data
284                tree.results = tree.rightBranch.results
285                tree.summary = tree.rightBranch.summary
286            tree.leftBranch = None
287            tree.rightBranch = None
288                
289                
290
CART算法实现
  1 '''
  2 Created on Oct 14, 2010
  3 
  4 @author: Peter Harrington
  5 '''
  6 import matplotlib.pyplot as plt
  7 
  8 decisionNode = dict(boxstyle="sawtooth", fc="0.8")
  9 leafNode = dict(boxstyle="circle", fc="0.7")
 10 arrow_args = dict(arrowstyle="<-")
 11 
 12 #获取树的叶子节点
 13 def getNumLeafs(myTree):
 14     numLeafs = 0
 15     #dict转化为list
 16     firstSides = list(myTree.keys())
 17     firstStr = firstSides[0]
 18     secondDict = myTree[firstStr]
 19     for key in secondDict.keys():
 20         #判断是否是叶子节点(通过类型判断,子类不存在,则类型为str;子类存在,则为dict)
 21         if type(secondDict[key]).__name__=='dict':#test to see if the nodes are dictonaires, if not they are leaf nodes
 22             numLeafs += getNumLeafs(secondDict[key])
 23         else:   numLeafs +=1
 24     return numLeafs
 25 
 26 #获取树的层数
 27 def getTreeDepth(myTree):
 28     maxDepth = 0
 29     #dict转化为list
 30     firstSides = list(myTree.keys())
 31     firstStr = firstSides[0]
 32     secondDict = myTree[firstStr]
 33     for key in secondDict.keys():
 34         if type(secondDict[key]).__name__=='dict':#test to see if the nodes are dictonaires, if not they are leaf nodes
 35             thisDepth = 1 + getTreeDepth(secondDict[key])
 36         else:   thisDepth = 1
 37         if thisDepth > maxDepth: maxDepth = thisDepth
 38     return maxDepth
 39 
 40 def plotNode(nodeTxt, centerPt, parentPt, nodeType):
 41     createPlot.ax1.annotate(nodeTxt, xy=parentPt,  xycoords='axes fraction',
 42              xytext=centerPt, textcoords='axes fraction',
 43              va="center", ha="center", bbox=nodeType, arrowprops=arrow_args )
 44     
 45 def plotMidText(cntrPt, parentPt, txtString):
 46     xMid = (parentPt[0]-cntrPt[0])/2.0 + cntrPt[0]
 47     yMid = (parentPt[1]-cntrPt[1])/2.0 + cntrPt[1]
 48     createPlot.ax1.text(xMid, yMid, txtString, va="center", ha="center", rotation=30)
 49 
 50 def plotTree(myTree, parentPt, nodeTxt):#if the first key tells you what feat was split on
 51     numLeafs = getNumLeafs(myTree)  #this determines the x width of this tree
 52     depth = getTreeDepth(myTree)
 53     firstSides = list(myTree.keys())
 54     firstStr = firstSides[0] #the text label for this node should be this         
 55     cntrPt = (plotTree.xOff + (1.0 + float(numLeafs))/2.0/plotTree.totalW, plotTree.yOff)
 56     plotMidText(cntrPt, parentPt, nodeTxt)
 57     plotNode(firstStr, cntrPt, parentPt, decisionNode)
 58     secondDict = myTree[firstStr]
 59     plotTree.yOff = plotTree.yOff - 1.0/plotTree.totalD
 60     for key in secondDict.keys():
 61         if type(secondDict[key]).__name__=='dict':#test to see if the nodes are dictonaires, if not they are leaf nodes   
 62             plotTree(secondDict[key],cntrPt,str(key))        #recursion
 63         else:   #it's a leaf node print the leaf node
 64             plotTree.xOff = plotTree.xOff + 1.0/plotTree.totalW
 65             plotNode(secondDict[key], (plotTree.xOff, plotTree.yOff), cntrPt, leafNode)
 66             plotMidText((plotTree.xOff, plotTree.yOff), cntrPt, str(key))
 67     plotTree.yOff = plotTree.yOff + 1.0/plotTree.totalD
 68 #if you do get a dictonary you know it's a tree, and the first element will be another dict
 69 #绘制决策树 样例1
 70 def createPlot(inTree):
 71     fig = plt.figure(1, facecolor='white')
 72     fig.clf()
 73     axprops = dict(xticks=[], yticks=[])
 74     createPlot.ax1 = plt.subplot(111, frameon=False, **axprops)    #no ticks
 75     #createPlot.ax1 = plt.subplot(111, frameon=False) #ticks for demo puropses 
 76     #宽,高间距
 77     plotTree.totalW = float(getNumLeafs(inTree))-3
 78     plotTree.totalD = float(getTreeDepth(inTree))-2
 79 #    plotTree.totalW = float(getNumLeafs(inTree))
 80 #    plotTree.totalD = float(getTreeDepth(inTree))
 81     plotTree.xOff = -0.5/plotTree.totalW; plotTree.yOff = 1.0;
 82     plotTree(inTree, (0.95,1.0), '')
 83     plt.show()
 84     
 85 #绘制决策树 样例2
 86 def createPlot1(inTree):
 87     fig = plt.figure(1, facecolor='white')
 88     fig.clf()
 89     axprops = dict(xticks=[], yticks=[])
 90     createPlot.ax1 = plt.subplot(111, frameon=False, **axprops)    #no ticks
 91     #createPlot.ax1 = plt.subplot(111, frameon=False) #ticks for demo puropses 
 92     #宽,高间距
 93     plotTree.totalW = float(getNumLeafs(inTree))-4.5
 94     plotTree.totalD = float(getTreeDepth(inTree)) -3
 95     plotTree.xOff = -0.5/plotTree.totalW; plotTree.yOff = 1.0;
 96     plotTree(inTree, (1.0,1.0), '')
 97     plt.show()
 98 
 99 #绘制树的根节点和叶子节点(根节点形状:长方形,叶子节点:椭圆形)
100 #def createPlot():
101 #    fig = plt.figure(1, facecolor='white')
102 #    fig.clf()
103 #    createPlot.ax1 = plt.subplot(111, frameon=False) #ticks for demo puropses 
104 #    plotNode('a decision node', (0.5, 0.1), (0.1, 0.5), decisionNode)
105 #    plotNode('a leaf node', (0.8, 0.1), (0.3, 0.8), leafNode)
106 #    plt.show()
107 
108 def retrieveTree(i):
109     listOfTrees =[{'no surfacing': {0: 'no', 1: {'flippers': {0: 'no', 1: 'yes'}}}},
110                   {'no surfacing': {0: 'no', 1: {'flippers': {0: {'head': {0: 'no', 1: 'yes'}}, 1: 'no'}}}}
111                   ]
112     return listOfTrees[i]
113 
114 #thisTree = retrieveTree(0)
115 #createPlot(thisTree)
116 #createPlot() 
117 #myTree = retrieveTree(0)
118 #numLeafs =getNumLeafs(myTree)
119 #treeDepth =getTreeDepth(myTree)
120 #print(u"叶子节点数目:%d"% numLeafs)
121 #print(u"树深度:%d"%treeDepth)
绘制决策树
 1 # -*- coding: utf-8 -*-
 2 """
 3 Created on Wed Aug 15 14:16:59 2018
 4 
 5 @author: weixw
 6 """
 7 import myCart as mc
 8 if __name__ == '__main__':
 9     import treePlotter as tp
10     dataSet = mc.loadCSV("dataSet.csv")
11     myTree = mc.createTree(dataSet, evaluationFunc=gini)
12     print(u"myTree:%s"%myTree)
13     #绘制决策树
14     print(u"绘制决策树:")
15     tp.createPlot1(myTree)
16     decisionTree = mc.buildDecisionTree(dataSet, evaluationFunc=gini)
17     testData = [5.9,3,4.2,1.75]
18     r = mc.classify(testData, decisionTree)
19     print(u"分类后测试结果:")
20     print(r)
21     print()
22     mc.prune(decisionTree, 0.4)   
23     r1 = mc.classify(testData, decisionTree)
24     print(u"剪枝后测试结果:")
25     print(r1)
测试代码

 

运行结果

为什么我要再写个createTree(dataSet, evaluationFunc=gini)函数,是因为绘制决策树createPlot1(myTree)输入参数需要是json结构数据。

 

将生成的决策树变为可视图形,这样更直观。

当然,也可以将自定义树对象信息打印出来,我在代码里已加入打印语句。

打印结果如下,因为屏幕的原因,没有全部粘贴出来,大家可以对照决策树绘制图,这样可以相互印证,加深理解。

 

 

在未做剪枝处理时的分类测试结果如下:

 

剪枝处理后的分类测试结果:

可以看出,{'versicolor': 47}取代了父结点serial:3,成为新的叶子结点。

 

参考文献

《统计学习方法》

《机器学习实战》

 

 

不要让懒惰占据你的大脑,不要让妥协拖垮你的人生。青春就是一张票,能不能赶上时代的快车,你的步伐掌握在你的脚下。

标签
易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!