朴素贝叶斯原理

讲的很好不赘述了：
https://zhuanlan.zhihu.com/p/26262151

朴素贝叶斯例子

朴素贝叶斯文本分类：

完整可运行代码在最底下！！！
要实现的任务是根据已有的训练集：

postinglist=[ ['my','dog','has','flea','problems','help','please'], ['maybe','not','take','him','to','dog','park','sutpid'], ['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'] ]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8

以及他的标签（0：表示积极 1：表示消极）：

classvec=[ 0, 1, 0, 1, 0, 1
]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8

进行处理（也可以说是训练吧）之后对我们输入的词：

testentry=['love','my','dalmation']
testentry=['stupid','garbage']

  

 

• 1
• 2

进行一个类别的判断。得到结果：

['love', 'my', 'dalmation'] is classified as:  0
['stupid', 'garbage'] is classified as:  1

  

 

• 1
• 2

具体过程

根据main函数中的testingNB介绍流程：

code1

def loaddataset(): postinglist=[ ['my','dog','has','flea','problems','help','please'], ['maybe','not','take','him','to','dog','park','sutpid'], ['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

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11

根据函数的定义

listposts=[ ['my','dog','has','flea','problems','help','please'], ['maybe','not','take','him','to','dog','park','sutpid'], ['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'] ]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8

listclasses=[0,1,0,1,0,1]

  

 

• 1

code2

def creatvocablist(dataset): vocabset=set([]) for document in dataset: vocabset=vocabset|set(document) return list(vocabset)

  

 

• 1
• 2
• 3
• 4
• 5

这个函数document依次为
[‘my’,‘dog’,‘has’,‘flea’,‘problems’,‘help’,‘please’]
[‘maybe’,‘not’,‘take’,‘him’,‘to’,‘dog’,‘park’,‘sutpid’]
[‘my’,‘dalmation’,‘is’,‘so’,‘cute’,‘i’,‘love’,‘him’]
…
set函数对列表去重，例如：

vocabset|set(document)是对两个列表取并集，例如：

因此这个函数执行完的输出为，对listposts去重后所有的单词的集合：

myvocablist = 
[
'ate', 'cute', 'help', 'sutpid', 'i', 'love', 'stop', 'worthless', 'how', 'him', 'so', 
'steak', 'my', 'food', 'quit', 'dog', 'garbage', 'flea', 'buying', 'please', 'take', 'has', 
'dalmation', 'stupid', 'not', 'licks', 'to', 'maybe', 'posting', 'park', 'problems', 'is', 
'mr'
]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7

这个集合也就是我们训练集的字典，长度33.

code3

这段代码的输入是：

myvocablist = 
[
'ate', 'cute', 'help', 'sutpid', 'i', 'love', 'stop', 'worthless', 'how', 'him', 'so', 
'steak', 'my', 'food', 'quit', 'dog', 'garbage', 'flea', 'buying', 'please', 'take', 'has', 
'dalmation', 'stupid', 'not', 'licks', 'to', 'maybe', 'posting', 'park', 'problems', 'is', 
'mr'
]

postinglist=[ ['my','dog','has','flea','problems','help','please'], ['maybe','not','take','him','to','dog','park','sutpid'], ['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'] ]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16

输出trainmat是：

字典长度是33，postinglist的长度是6也就是6条训练数据，输出的trainmat长度是6 * 33
其实这段代码就是对输入的文本根据我们的字典进行编码：
例如第一行：
[0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0]
根据字典：

myvocablist = 
[
'ate', 'cute', 'help', 'sutpid', 'i', 'love', 'stop', 'worthless', 'how', 'him', 'so', 
'steak', 'my', 'food', 'quit', 'dog', 'garbage', 'flea', 'buying', 'please', 'take', 'has', 
'dalmation', 'stupid', 'not', 'licks', 'to', 'maybe', 'posting', 'park', 'problems', 'is', 
'mr'
]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7

可以解码成：help my dog flea please has problems
与第一条语料对应：

code4

def trainNB0(trainmatrix,traincategory): numtraindocs=len(trainmatrix) numwords=len(trainmatrix[0]) pAbusive=sum(traincategory)/float(numtraindocs) p0num=np.ones(numwords);p1num=np.ones(numwords) p0denom=2.0;p1denom=2.0 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=np.log(p1num/float(p1denom)) p0vec=np.log(p0num/float(p0denom)) return p0vec,p1vec,pAbusive

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16

我们给这个函数的输入为编码的语料和这个语料对应的标签：

[
0,
1,
0,
1,
0,
1
]

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8

因此可以知道每个值：

def trainNB0(trainmatrix,traincategory): numtraindocs=len(trainmatrix) # 6 numwords=len(trainmatrix[0]) # 33 pAbusive=sum(traincategory)/float(numtraindocs) # (0 + 1 + 0 + 1 + 0 + 1) / 6 = 1/2 p0num=np.ones(numwords);p1num=np.ones(numwords) # 创建两个33个1的矩阵, 拉普拉斯平滑（加1平滑），解决0概率出现问题 p0denom=2.0;p1denom=2.0

  

 

• 1
• 2
• 3
• 4
• 5
• 6

下面这段代码就是判断当前语料的标签是0还是1，是1的把这个语料中每个词出现几次加到p1num中存起来。是0的加到p0num中。p1denom记录1类别中总的词数，p0denom同理。

 for i in range(numtraindocs): if traincategory[i]==1: p1num+=trainmatrix[i] p1denom+=sum(trainmatrix[i]) else: p0num+=trainmatrix[i] p0denom+=sum(trainmatrix[i])

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7

p1vec=np.log(p1num/float(p1denom))#消极类中该词汇出现的概率，取对数是防止浮点数下溢出
p0vec=np.log(p0num/float(p0denom))#积极类中该词汇的概率

  

 

• 1
• 2

浮点数下溢解释：当两个非常小的数相乘时，所得结果会更小，计算机在处理浮点数时，如果超出了浮点数的表示范围就会丢失掉非常低位的精度，由于log有性质：ln(a * b) = ln(a) + ln(b)，因此可以对连乘进行取log这样换成log值相加就不会出现浮点数下溢的情况。

code5

def setofwords2vec(vocabset,inputset): returnvec=[0]*len(vocabset) for word in inputset: if word in vocabset: returnvec[vocabset.index(word)]=1 else: print("the word %s is not in my vocabulary"%word) return returnvec

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8

这段代码和处理训练的语料一样，把用于测试的语料转成1 * 33的向量表示。

code6

def classifyNB(vec2Classify,p0vec,p1vec,pclass1): p1=sum(vec2Classify*p1vec)+np.log(pclass1) p0=sum(vec2Classify*p0vec)+np.log(1-pclass1) if p1>p0: return 1 else: return 0

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7

p1=sum(vec2Classifyp1vec)+np.log(pclass1)这句把标签为1且测试的单词在训练集中出现的频率累加
p0=sum(vec2Classifyp0vec)+np.log(1-pclass1)这句把标签为0且测试的单词在训练集中出现的频率累加
最后判断p1和p0谁大就属于哪一个类别！

完整代码(python3可运行)

import numpy as np

def loaddataset(): postinglist=[ ['my','dog','has','flea','problems','help','please'], ['maybe','not','take','him','to','dog','park','sutpid'], ['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


def creatvocablist(dataset): vocabset=set([]) for document in dataset: vocabset=vocabset|set(document) return list(vocabset)

def setofwords2vec(vocabset,inputset): returnvec=[0]*len(vocabset) for word in inputset: if word in vocabset: returnvec[vocabset.index(word)]=1 else: print("the word %s is not in my vocabulary"%word) return returnvec

def trainNB0(trainmatrix,traincategory): numtraindocs=len(trainmatrix) numwords=len(trainmatrix[0]) pAbusive=sum(traincategory)/float(numtraindocs) p0num=np.ones(numwords);p1num=np.ones(numwords) p0denom=2.0;p1denom=2.0 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=np.log(p1num/float(p1denom)) p0vec=np.log(p0num/float(p0denom)) return p0vec,p1vec,pAbusive

def classifyNB(vec2Classify,p0vec,p1vec,pclass1): p1=sum(vec2Classify*p1vec)+np.log(pclass1) p0=sum(vec2Classify*p0vec)+np.log(1-pclass1) if p1>p0: return 1 else: return 0


def testingNB(): listposts,listclasses=loaddataset() myvocablist=creatvocablist(listposts) trainmat=[] for words in listposts: trainmat.append(setofwords2vec(myvocablist,words)) p0vec,p1vec,pab=trainNB0(np.array(trainmat),np.array(listclasses)) testentry=['love','my','dalmation'] vec2Classify=np.array(setofwords2vec(myvocablist,testentry)) print(testentry,"is classified as: ",classifyNB(vec2Classify,p0vec,p1vec,pab)) testentry=['stupid','garbage'] vec2Classify=np.array(setofwords2vec(myvocablist,testentry)) print(testentry,"is classified as: ",classifyNB(vec2Classify,p0vec,p1vec,pab))


if __name__ == '__main__': testingNB()

  

 

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30
• 31
• 32
• 33
• 34
• 35
• 36
• 37
• 38
• 39
• 40
• 41
• 42
• 43
• 44
• 45
• 46
• 47
• 48
• 49
• 50
• 51
• 52
• 53
• 54
• 55
• 56
• 57
• 58
• 59
• 60
• 61
• 62
• 63
• 64
• 65
• 66
• 67
• 68
• 69
• 70
• 71
• 72
• 73

运行结果：

文章来源: blog.csdn.net，作者：只会git clone的程序员，版权归原作者所有，如需转载，请联系作者。

原文链接：blog.csdn.net/qq_37668436/article/details/115034088