贝叶斯文本分类--python代码

from numpy import *

def loadDataSet():
    postingList=[['my', 'dog', 'has', 'flea', 'problems', 'help', 'please'],
                 ['maybe', 'not', 'take', 'him', 'to', 'dog', 'park', 'stupid'],
                 ['my', 'dalmation', 'is', 'so', 'cute', 'I', 'love', 'him'],
                 ['stop', 'posting', 'stupid', 'worthless', 'garbage'],
                 ['mr', 'licks', 'ate', 'my', 'steak', 'how', 'to', 'stop', 'him'],
                 ['quit', 'buying', 'worthless', 'dog', 'food', 'stupid']]
    classVec = [0,1,0,1,0,1]    #1 is abusive, 0 not
    return postingList,classVec
    #生成的list是二位数组，访问时候应该加list[0],list[1],list[2]

def createVocabList(dataSet):
    vocabSet = set([])  #create empty set，set集合中不会重复
    for document in dataSet:
        vocabSet = vocabSet | set(document) #union of the two sets
        #   |  求两者的并集
    return list(vocabSet)  #list去除重复词

def setOfWords2Vec(vocabList, inputSet):
    #前者是词汇表，后者是输入的文档
    returnVec = [0]*len(vocabList)  #生成len长度的0的集合
    for word in inputSet:
        if word in vocabList:
            returnVec[vocabList.index(word)] = 1 #文本数量+1
        else: print("the word: %s is not in my Vocabulary!" % word)
    return returnVec

def bagOfWords2VecMN(vocabList, inputSet):
    returnVec = [0]*len(vocabList)
    for word in inputSet:
        if word in vocabList:
            returnVec[vocabList.index(word)] += 1  #文本数量+1，频率增大
    return returnVec

def trainNB0(trainMatrix,trainCategory):#导入文档矩阵和文档标签
    numTrainDocs = len(trainMatrix)   #文档是二维矩阵，不同文档/行，文档中不同词汇/列
    numWords = len(trainMatrix[0])
    pAbusive = sum(trainCategory)/float(numTrainDocs)#标签计算，带有侮辱性词汇+1，除以文档数行，带侮辱的概率
    p0Num = ones(numWords); p1Num = ones(numWords)  #change to ones() ，列/词数置为全1向量，防止概率为零
    p0Denom = 2.0; p1Denom = 2.0                        #change to 2.0
    for i in range(numTrainDocs):   #遍历行
        if trainCategory[i] == 1:  #1 侮辱
            p1Num += trainMatrix[i]     #所有侮辱文章listj    listj[侮辱1]+listj[侮辱2]+listj[侮辱3]
            print(p1Num)
            p1Denom += sum(trainMatrix[i]) #带侮辱性，list[侮辱总]  出现次数之和
            print(p1Denom)
           #6篇文档，0.5概率，运行了3次
        else:
            p0Num += trainMatrix[i] #
            p0Denom += sum(trainMatrix[i])
    p1Vect = log(p1Num/p1Denom)        #change to log() 以e为底自然对数
    p0Vect = log(p0Num/p0Denom)         #change to log()
    return p0Vect,p1Vect,pAbusive#前者p(有该词|是侮辱文)，后者p(有该词|是正常文)，p(侮辱文)
#p(侮辱问|有该词)=p(有该词|是侮辱文)*p(侮辱文)/p(有该词)
#p(有该词|是侮辱)=p(侮辱文|有该词)*p(有该词)/p(侮辱文)


def classifyNB(vec2Classify, p0Vec, p1Vec, pClass1):
    p1 = sum(vec2Classify * p1Vec)-log(pClass1)    #element-wise mult
    p0 = sum(vec2Classify * p0Vec)-log(1.0 - pClass1) #对应单词乘以对应概率，
    #为什么是+？
    if p1 > p0:
        return 1
    else:
        return 0

def testingNB():
    listOPosts,listClasses = loadDataSet()  #建立文档行/单词列，标签列
    myVocabList = createVocabList(listOPosts) #所有单词去重，建立列表
    trainMat=[]
    for postinDoc in listOPosts:
        trainMat.append(bagOfWords2VecMN(myVocabList, postinDoc)) #双遍历，每篇单词中在所有单词中出现次数
    p0V,p1V,pAb = trainNB0(array(trainMat),array(listClasses))#前者p(有该词|是侮辱文)，后者p(有该词|是正常文)，p(侮辱文)
    testEntry = ['love', 'my', 'dalmation']
    thisDoc = array(bagOfWords2VecMN(myVocabList, testEntry))#test在总单词中出现次数
    print(testEntry,'classified as: ',classifyNB(thisDoc,p0V,p1V,pAb))
    testEntry = ['stupid', 'garbage']
    thisDoc = array(setOfWords2Vec(myVocabList, testEntry))
    print(testEntry,'classified as: ',classifyNB(thisDoc,p0V,p1V,pAb))

testingNB()
#A 中有2个灰球，2个黑球；B 中有1个灰球，2个黑球
#A|gray=1/2     B/gray=1/3
#A|black=1/2    B/black=2/3
#p(gray|B)=p(gray)*p(B|gray)/p(B)=(3/7)*(1/3)/(3/7)=1/3
#p(有该词|垃圾邮件)=p(垃圾邮件|有该词)*p(有该词)/p(垃圾邮件)
#                  有0.5概率是垃圾邮件*（训练的词*垃圾邮件中的词）
#A 垃圾邮件   B好邮件   灰球-好词  黑球-坏词  无法预先判断好词还是坏词