pythonde 朴素贝叶斯算法

#coding=utf-8

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

#创建一个带有所有单词的列表

def createVocabList(dataSet):

    vocabSet = set([])

    for document in dataSet:

        vocabSet = vocabSet | set(document)

    return list(vocabSet)

def setOfWords2Vec(vocabList, inputSet):

    retVocabList = [0] * len(vocabList)

    for word in inputSet:

        if word in vocabList:

            retVocabList[vocabList.index(word)] = 1

        else:

            print 'word ',word ,'not in dict'

    return retVocabList

#另一种模型   

def bagOfWords2VecMN(vocabList, inputSet):

    returnVec = [0]*len(vocabList)

    for word in inputSet:

        if word in vocabList:

            returnVec[vocabList.index(word)] += 1

    return returnVec

def trainNB0(trainMatrix,trainCatergory):

    numTrainDoc = len(trainMatrix)

    numWords = len(trainMatrix[0])

    pAbusive = sum(trainCatergory)/float(numTrainDoc)

    #防止多个概率的成绩当中的一个为0

    p0Num = ones(numWords)

    p1Num = ones(numWords)

    p0Denom = 2.0

    p1Denom = 2.0

    for i in range(numTrainDoc):

        if trainCatergory[i] == 1:

            p1Num +=trainMatrix[i]

            p1Denom += sum(trainMatrix[i])

        else:

            p0Num +=trainMatrix[i]

            p0Denom += sum(trainMatrix[i])

    p1Vect = log(p1Num/p1Denom)#处于精度的考虑，否则很可能到限归零

    p0Vect = log(p0Num/p0Denom)

    return p0Vect,p1Vect,pAbusive

def classifyNB(vec2Classify, p0Vec, p1Vec, pClass1):

    p1 = sum(vec2Classify * p1Vec) + log(pClass1)    #element-wise mult

    p0 = sum(vec2Classify * p0Vec) + log(1.0 - pClass1)

    if p1 > p0:

        return 1

    else:

        return 0

def testingNB():

    listOPosts,listClasses = loadDataSet()

    myVocabList = createVocabList(listOPosts)

    trainMat=[]

    for postinDoc in listOPosts:

        trainMat.append(setOfWords2Vec(myVocabList, postinDoc))

    p0V,p1V,pAb = trainNB0(array(trainMat),array(listClasses))

    testEntry = ['love', 'my', 'dalmation']

    thisDoc = array(setOfWords2Vec(myVocabList, testEntry))

    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)

def main():

    testingNB()

if __name__ == '__main__':

    main()