【python与机器学习入门3】朴素贝叶斯2——垃圾邮件分类

参考博客：朴素贝叶斯基础篇之言论过滤器 （po主Jack-Cui,《——大部分内容转载自

                 

参考书籍：《机器学习实战》——第四章4.6

朴素贝叶斯基础内容见前篇《——

1 数据集

ham文件夹 正常邮件

spam文件夹 垃圾邮件

2 数据处理

3 训练

4 测试

#!/usr/bin/env python
#_*_coding:utf-8_*_
from numpy import *
import re
'''
    文本分类-过滤垃圾邮件
'''


'''切分文本'''
def textParse(bigString):
    listOfTokens = re.split(r'\W*',bigString)
    return [tok.lower() for tok in listOfTokens if len(tok)>2]


'''创建词汇表'''
def createVocablist(dataset):
    vocabList = set([])
    for data in dataset:
        vocabList = vocabList | set(data)
    return list(vocabList)

'''词汇转向量'''
def setOfWord2Vec(vocabList,dataset):
    returnVec = [0] * len(vocabList)
    for word in dataset:
        if word in vocabList:
            returnVec[vocabList.index(word)] = 1 #单词出现标1
        else:
            print "sorry this word %s is not in our vocablist" % word
    return  returnVec


'''计算p(wi|ci)
   该类下，该单词出现的概率
   需要计算每一类每个单词出现的次数（分子） 和 每一类出现过的总单词数（分母）
'''
def trainNB(trainMatrix,trainCategory):
    numTrainDocs = len(trainMatrix)
    numWords = len(trainMatrix[0])
    pSpam = float(sum(trainCategory)) / float(numTrainDocs)
    p1Num = ones(numWords)
    p0Num = ones(numWords) #记录每个单词在该类出现的次数，是一个向量
    p1Denom = 2.0
    p0Denom = 2.0 #属于该类的总单词出现次数,是一个数
    for i in range(numTrainDocs):
        if trainCategory[i] == 1:
            p1Num += trainMatrix[i]
            p1Denom += sum(trainMatrix[i])
        else:
            p0Num += trainMatrix[i]
            p0Denom += sum(trainMatrix[i])
    p1Vec = log(p1Num / p1Denom)
    p0Vec = log(p0Num / p0Denom)
    return p0Vec,p1Vec,pSpam

'''计算文档属于某个类别i的概率：p = p(w0|ci)p(w1|ci)p(w2|ci)~p(wn|ci)*p（ci）以下是二分类'''
def classifyNB(vec2classify,p0vec,p1vec,pclass1):
    p1 = sum(vec2classify * p1vec) + log(pclass1)
    p0 = sum(vec2classify * p0vec) + log(1 - pclass1)
    if p1 > p0:
        #print "p1=%f" % p1
        return 1
    else:
        #print "p0=%f" % p0
        return 0


'''
    
    random.uniform(a, b)，用于生成一个指定范围内的随机符点数
'''

def spamTest():
    classList = [] ; docList = []
    '''导入文件共50个'''
    for i in range(1,26):
        wordList = textParse(open('email/spam/%d.txt' % i,'r').read())
        classList.append(1)
        docList.append(wordList)
        wordList = textParse(open('email/ham/%d.txt' % i, 'r').read())
        classList.append(0)
        docList.append(wordList)
    '''创建词汇表'''
    vocabList = createVocablist(docList)
    #print len(vocabList)
    #print vocabList
    trainSetIndex = range(50)
    testSetIndex = []
    #print docList
    '''划分训练集和测试集 取10个测试集 40个训练集'''
    for i in range(10):
        randIndex = int(random.uniform(0,len(trainSetIndex)))
        testSetIndex.append(trainSetIndex[randIndex])#避免重复
        #print  trainSetIndex[randIndex]
        del(trainSetIndex[randIndex])
    numTrainDoc = len(trainSetIndex)
    numTestDoc = len(testSetIndex)
    trainMat = [];  testMat = [];
    trainClass = []; testClass = [];
    '''训练集数据向量化 数据集标签'''
    for i in range(numTrainDoc):
        trainMat.append(setOfWord2Vec(vocabList,docList[trainSetIndex[i]]))
        trainClass.append(classList[trainSetIndex[i]])
    p0Vec, p1Vec, pSpam = trainNB(trainMat,trainClass)
    errorCount = 0.0
    '''测试集数据向量化 数据集标签 进行测试'''
    for i in range(numTestDoc):
        testMat = setOfWord2Vec(vocabList,docList[testSetIndex[i]])
        testClass = classList[testSetIndex[i]]
        classResult = classifyNB(testMat,p0Vec,p1Vec,pSpam)
        if classResult != testClass:
            print "classify wrong：origin %d" % testClass
            errorCount += 1
    print "error rate = %.2f" % (errorCount / numTestDoc)



if __name__ == '__main__':
   spamTest()