算法优缺点
优点:在数据较少的情况下依然有效,可以处理多类别问题
缺点:对输入数据的准备方式敏感
适用数据类型:标称型数据
算法思想:
朴素贝叶斯
比如我们想判断一个邮件是不是垃圾邮件,那么我们知道的是这个邮件中的词的分布,那么我们还要知道:垃圾邮件中某些词的出现是多少,就可以利用贝叶斯定理得到。
朴素贝叶斯分类器中的一个假设是:每个特征同等重要
贝叶斯分类是一类分类算法的总称,这类算法均以贝叶斯定理为基础,故统称为贝叶斯分类。
loadDataSet()
创建数据集,这里的数据集是已经拆分好的单词组成的句子,表示的是某论坛的用户评论,标签1表示这个是骂人的
createVocabList(dataSet)
找出这些句子中总共有多少单词,以确定我们词向量的大小
setOfWords2Vec(vocabList,inputSet)
将句子根据其中的单词转成向量,这里用的是伯努利模型,即只考虑这个单词是否存在
bagOfWords2VecMN(vocabList,inputSet)
这个是将句子转成向量的另一种模型,多项式模型,考虑某个词的出现次数
trainNB0(trainMatrix,trainCatergory)
计算P(i)和P(w[i]|C[1])和P(w[i]|C[0]),这里有两个技巧,一个是开始的分子分母没有全部初始化为0是为了防止其中一个的概率为0导致整体为0,另一个是后面乘用对数防止因为精度问题结果为0
classifyNB(vec2Classify,p0Vec,p1Vec,pClass1)
根据贝叶斯公式计算这个向量属于两个集合中哪个的概率高
# @param 朴素贝叶斯算法的python实现
# @author 编程之家 jb51.cc|www.512Pic.com
#coding=utf-8
from numpy import *
def loadDataSet():
postingList=[['my','dog','has','flea','problems','help','please'],['maybe','not','take','him','to','park','stupid'],['my','dalmation','is','so','cute','I','love','him'],['stop','posting','stupid','worthless','garbage'],['mr','licks','ate','my','steak','how','stop',['quit','buying','food','stupid']]
classVec = [0,1,1] #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,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','dalmation']
thisDoc = array(setOfWords2Vec(myVocabList,testEntry))
print testEntry,'classified as: ',classifyNB(thisDoc,p0V,pAb)
testEntry = ['stupid','garbage']
thisDoc = array(setOfWords2Vec(myVocabList,pAb)
def main():
testingNB()
if __name__ == '__main__':
main()
# End www.jb51.cc