机器学习经典算法2-决策树

本文链接：https://blog.csdn.net/moodytong/article/details/9613327

本文介绍了决策树的基本理念，强调其相比knn算法在数据理解上的优势。通过信息增益和熵来构建决策树，并概述了算法的一般流程，包括数据收集、准备、分析、训练、测试和使用。此外，还提到了决策树构造的终止条件和伪代码。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

一、算法简要

决策树的基本理念就是通过不断的条件筛选，从而得到最后的答案。knn算法最大的缺点就是无法给出数据的内在含义，而决策树则在数据形式非常容易理解，有一定的实际意义。

这里所讲到的决策树非叶子节点的建立是依据信息增益和熵的概念，这个可以自己去查。通过计算按特定属性划分数据集前后发生的熵的变化，选择信息增益最大的特征属性作为分叉节点，从而一步一步进行决策树的构造。

二、算法一般流程

1.收集数据：任意方法和途径

2.准备数据：数据必须离散化

3.分析数据：构造树完成后，检查图形是否符合预测

4.训练算法：决策树的构造

5.测试算法：一般将决策树用于分类，可以用错误率衡量

6.使用算法：决策树可以用于任何监督学习算法

三、算法伪代码

1.判断决策树是否构造完毕，否则执行2，是则执行3。有终止条件：所有的类标签完全相同，返回该类标签即可；使用完了所有的特征，仍然不能将数据集划分成包含唯一类别的分组，此时，返回出现次数较多的类别

2.选择时的信息增益最大的特征作为分叉节点，对分叉节点的每个特定值，调用1

3.决策树构造完毕，对测试数据进行分类

四、代码实现与示例

import matplotlib.pyplot as plt
from math import log
import operator
'''calculate shannon entropy:count different distribution first then calculate'''
def calcShannonEnt(dataSet):
    numEntries = len(dataSet)
    labelCounts={}
    for featVec in dataSet:
        currentLabel = featVec[-1]
        if currentLabel not in labelCounts.keys():
            labelCounts[currentLabel]=0
        labelCounts[currentLabel]+=1
    shannonEnt = 0.0
    for key in labelCounts:
        prob = float(labelCounts[key])/numEntries
        shannonEnt -= prob*log(prob,2)
    return shannonEnt
'''split the dataSet with the given feature index, form new dataSet without the given feature'''
def splitDataSet(dataSet, axis, value):
    retDataSet = []
    for featVec in dataSet:
        if featVec[axis] == value:
            reducedFeatVec = featVec[:axis]
            reducedFeatVec.extend(featVec[axis+1:])
            retDataSet.append(reducedFeatVec)
    return retDataSet
'''choose the besf feature index with the largest information gain'''
def chosenBestFeatureToSplit(dataSet):
    numFeatures = len(dataSet[0])-1
    baseEntropy = calcShannonEnt(dataSet)
    bestInfoGain = 0
    bestFeature = -1
    for i in range(numFeatures):
        featList = [example[i] for example in dataSet]
        uniqueVals = set(featList)
        newEntropy = 0.0
        for value in uniqueVals:
            subDataSet = splitDataSet(dataSet, i, value)
            prob = len(subDataSet)/float(len(dataSet))
            newEntropy += prob*calcShannonEnt(subDataSet)
        infoGain = baseEntropy - newEntropy
        if(infoGain > bestInfoGain):
            bestInfoGain = newEntropy
            bestFeature = i
    return bestFeature
'''return the majority of the list, for figure out the class with unpurity examples'''
def majorityCnt(classList):
    classCount = {}
    for vote in classList:
        if vote not in classCount.keys():
            classCount[vote] = 0
        classCount[vote] += 1
    sortedClassCount = sorted(classCount.iteritems(), key=operator.itemgetter(1),
                              reverse=True)
    return sortedClassCount[0][0]
def createTree(dataSet, labels):
    '''if we modify labels_t, python will not labels simutaneously'''
    labels_t = labels[:]
    classList = [example[-1] for example in dataSet]
    if classList.count(classList[0]) == len(classList):
        return classList[0]
    if len(dataSet[0])==1:
        return majorityCnt(classList)
    bestFeat = chosenBestFeatureToSplit(dataSet)
    bestFeatLabel = labels[bestFeat]
    myTree={bestFeatLabel:{}}
    del(labels_t[bestFeat])
    featValues =[example[bestFeat] for example in dataSet]
    uniqueVals = set(featValues)
    for value in uniqueVals:
        subLabels = labels_t[:]
        myTree[bestFeatLabel][value] = createTree(splitDataSet(dataSet, bestFeat, value),subLabels)
    return myTree
def storeTree(inputTree, filename):
    import pickle
    fw = open(filename,'w')
    pickle.dump(inputTree,fw)
    fw.close()
def grabTree(filename):
    import pickle
    fr = open(filename)
    return pickle.load(fr)
def classify(inputTree, featLabels, testVec):
    firstStr = inputTree.keys()[0]
    secondDict = inputTree[firstStr]
    featIndex = featLabels.index(firstStr)
    for key in secondDict.keys():
        if testVec[featIndex]==key:
            if type(secondDict[key]).__name__=='dict':
                classLabel = classify(secondDict[key], featLabels, testVec)
            else:
                classLabel = secondDict[key]
    return classLabel
myDat = [[1,1,'yes'],
         [1,1,'yes'],
         [1,0,'no'],
         [0,1,'no'],
         [0,1,'no']
         ]
labels = ['no surfacing','flippers']
myT = createTree(myDat, labels)
print myT
fileName="tmpTree.txt"
storeTree(myT,fileName)
print grabTree(fileName)
print classify(myT, labels, [1,0])