以在线社区的留言板为例。为了不影响社区的发展,我们要屏蔽侮辱性的言论,所以要构建一个快速过滤器,如果某条留言使用了负面或者侮辱性的语言,那么就将该留言标识为内容不当。过滤这类内容是一个很常见的需求。对此问题建立两个类别:侮辱类和非侮辱类,使用1和0分别标识。
有以下先验数据,使用bayes算法对未知类别数据分类:
帖子内容 | 类别 |
---|---|
‘my’,‘dog’,‘has’,‘flea’,‘problems’,‘help’,‘please’ | 0 |
‘maybe’,‘not’,‘take’,‘him’,‘to’,‘dog’,‘park’,'stupid | 1 |
‘my’,‘dalmation’,‘is’,‘so’,‘cute’,‘I’,‘love’,‘him’ | 0 |
‘stop’,‘posting’,‘stupid’,‘worthless’,'garbage | 1 |
‘mr’,‘licks’,‘ate’,‘my’,‘steak’,‘how’,‘to’,‘stop’,‘him’ | 0 |
‘quit’,‘buying’,‘worthless’,‘dog’,‘food’,‘stupid’ | 1 |
待分类数据:
关键字 | 类别 |
---|---|
‘love’,‘my’,‘dalmation’ | ? |
‘stupid’,‘garbage’ | ? |
参见上一节 05 机器学习 - 朴素贝叶斯分类算法原理
(1) 词表到词向量的转换函数
from numpy import * #过滤网站的恶意留言 # 创建一个实验样本 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] return postingList, classVec # 创建一个包含在所有文档中出现的不重复词的列表 def createVocabList(dataSet): vocabSet = set([]) #创建一个空集 for document in dataSet: vocabSet = vocabSet | set(document) #创建两个集合的并集 return list(vocabSet) #将文档词条转换成词向量 def setOfWords2Vec(vocabList, inputSet): returnVec = [0]*len(vocabList) #创建一个其中所含元素都为0的向量 for word in inputSet: if word in vocabList: #returnVec[vocabList.index(word)] = 1 #index函数在字符串里找到字符第一次出现的位置 词集模型 returnVec[vocabList.index(word)] += 1 #文档的词袋模型 每个单词可以出现多次 else: print "the word: %s is not in my Vocabulary!" % word return returnVec
(2)从词向量计算概率
#朴素贝叶斯分类器训练函数 从词向量计算概率 def trainNB0(trainMatrix, trainCategory): numTrainDocs = len(trainMatrix) numWords = len(trainMatrix[0]) pAbusive = sum(trainCategory)/float(numTrainDocs) # p0Num = zeros(numWords); p1Num = zeros(numWords) #p0Denom = 0.0; p1Denom = 0.0 p0Num = ones(numWords); #避免一个概率值为0,最后的乘积也为0 p1Num = ones(numWords); #用来统计两类数据中,各词的词频 p0Denom = 2.0; #用于统计0类中的总数 p1Denom = 2.0 #用于统计1类中的总数 for i in range(numTrainDocs): if trainCategory[i] == 1: p1Num += trainMatrix[i] p1Denom += sum(trainMatrix[i]) else: p0Num += trainMatrix[i] p0Denom += sum(trainMatrix[i]) # p1Vect = p1Num / p1Denom #p0Vect = p0Num / p0Denom p1Vect = log(p1Num / p1Denom) #在类1中,每个次的发生概率 p0Vect = log(p0Num / p0Denom) #避免下溢出或者浮点数舍入导致的错误 下溢出是由太多很小的数相乘得到的 return p0Vect, p1Vect, pAbusive
(3)根据现实情况修改分类器
注意:主要从以下两点对分类器进行修改
#朴素贝叶斯分类器 def classifyNB(vec2Classify, p0Vec, p1Vec, pClass1): p1 = sum(vec2Classify*p1Vec) + log(pClass1) 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)
(4)运行测试
>>>reload(bayes) <module ‘bayes’ from ‘bayes.py’> >>>bayes.testingNB() ['love','my','dalmation'] classified as: 0 ['stupid','garbage'] classified as: 1