机器学习实战 朴素贝叶斯

朴素贝叶斯

朴素贝叶斯概述

文本分类

准备数据:从文-本中构建词向量

-

训练算法：从词向量计算概率

-

贝叶斯分类函数

import numpy as np 
import matplotlib.pyplot as plt 
from numpy import *
"""
function:
    创建数据集
parameters:
    无
returns:
    postingList - 数据集
    classVec - 标签集
"""

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 

"""
function:
    从数据集中提取词汇表(不重复)
parameters:
    dataSet - 数据集
retunrns:
    vocalSet - 不重复的词汇表
"""

def createVocabList(dataSet):
    vocabSet = set([])
    for document in dataSet:
        vocabSet = vocabSet | set(document)
    return list(vocabSet)

"""
function:
    根据之前创建的词汇表来对输入数据进行向量化
parameters:
    vocabList - 词汇表
    inputSet - 输入的一个文档
returns:
    returnVec - 该文档向量
"""

def setOfWords2Vec(vocabList, inputSet):
    returnVec = [0]*len(vocabList)
    for word in inputSet:
        if word in vocabList:
            returnVec[vocabList.index(word)] = 1
        else:#前面创建词汇表有问题
            print("the word: %s is not in my Vocalbulary" % word)
    return returnVec 

"""
function:
    朴素贝叶斯分类器训练函数
parameters:
    trainMatrix - 训练文档矩阵,每篇文档调用set0OfWord2Vec生成的returnVec组成的矩阵
    trainCategory - 标签向量
returns:
    p0Vec - 侮辱类的条件概率数组，即每个词汇属于侮辱类的概率
    p1Vect - 非侮辱类的条件概率数组，即每个词汇属于非侮辱类的概率
    pAbusive - 文档属于侮辱类的概率
"""

def trainNB0(trainMatrix, trainCategory):
    numTrainDocs = len(trainMatrix)#文档数目
    numWords = len(trainMatrix[0])#每篇文档的词条数
    pAbusive = sum(trainCategory)/float(numTrainDocs)#文档属于侮辱类的概率
    #词条出现数初始化
    p0Num = ones(numWords)
    p1Num = ones(numWords)
    #分母初始化
    p0Denom = 2.0
    p1Denom = 2.0
    for i in range(numTrainDocs):#对每篇文档
        if trainCategory[i] == 1:#统计侮辱类的条件概率
            p1Num += trainMatrix[i]
            p1Denom += sum(trainMatrix[i])#这里我是理解成全概率公式
        else:
            p0Num += trainMatrix[i]
            p0Denom += sum(trainMatrix[i])
    p0Vect = log(p0Num/p0Denom)
    p1Vect = log(p1Num/p1Denom) 
    return p0Vect, p1Vect, pAbusive 

"""
function:
    朴素贝叶斯分类函数
parameters:
    vec2Classify - 待分类向量
    p0Vec - 侮辱类的条件概率数组
    p1Vec - 非侮辱类的条件概率数组
    pClass1 - 文档属于侮辱类的概率
returns:

"""

def classifyNB(vec2Classify, p0Vec, p1Vec, pClass1):
    #对数相加，相当与计算两个向量相乘的结果
    #这里不太懂，感觉和公式对不上
    p1 = sum(vec2Classify * p1Vec) + log(pClass1)
    p0 = sum(vec2Classify * p0Vec) + log(1.0 - pClass1)
    if p1 > p0:
        return 1
    else: 
        return 0

if __name__ == "__main__":
    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))

词袋模型

一个小优化，相比与之前只统计词出现与否的词条模型，词袋模型统计词出现的次数

-

垃圾邮件过滤

不清楚为什么我做出来的错误率这么高，算了，先放着吧

import re 
import random 
from array import *
import numpy as np
from numpy import *

"""
function:
    从数据集中提取词汇表(不重复)
parameters:
    dataSet - 数据集
retunrns:
    vocalSet - 不重复的词汇表
"""

def createVocabList(dataSet):
    vocabSet = set([])
    for document in dataSet:
        vocabSet = vocabSet | set(document)
    return list(vocabSet)

"""
function:
    根据之前创建的词汇表来对输入数据进行向量化
parameters:
    vocabList - 词汇表
    inputSet - 输入的一个文档
returns:
    returnVec - 该文档向量
"""

def setOfWords2Vec(vocabList, inputSet):
    returnVec = [0]*len(vocabList)
    for word in inputSet:
        if word in vocabList:
            returnVec[vocabList.index(word)] = 1
        else:#前面创建词汇表有问题
            print("the word: %s is not in my Vocalbulary" % word)
    return returnVec 

"""
function:
    朴素贝叶斯分类器训练函数
parameters:
    trainMatrix - 训练文档矩阵,每篇文档调用set0OfWord2Vec生成的returnVec组成的矩阵
    trainCategory - 标签向量
returns:
    p0Vec - 侮辱类的条件概率数组，即每个词汇属于侮辱类的概率
    p1Vect - 非侮辱类的条件概率数组，即每个词汇属于非侮辱类的概率
    pAbusive - 文档属于侮辱类的概率
"""

def trainNB0(trainMatrix, trainCategory):
    numTrainDocs = len(trainMatrix)#文档数目
    numWords = len(trainMatrix[0])#每篇文档的词条数
    pAbusive = sum(trainCategory)/float(numTrainDocs)#文档属于侮辱类的概率
    #词条出现数初始化
    p0Num = ones(numWords)
    p1Num = ones(numWords)
    #分母初始化
    p0Denom = 2.0
    p1Denom = 2.0
    for i in range(numTrainDocs):#对每篇文档
        if trainCategory[i] == 1:#统计侮辱类的条件概率
            p1Num += trainMatrix[i]
            p1Denom += sum(trainMatrix[i])#这里我是理解成全概率公式
        else:
            p0Num += trainMatrix[i]
            p0Denom += sum(trainMatrix[i])
    p0Vect = log(p0Num/p0Denom)
    p1Vect = log(p1Num/p1Denom) 
    return p0Vect, p1Vect, pAbusive 

"""
function:
    朴素贝叶斯分类函数
parameters:
    vec2Classify - 待分类向量
    p0Vec - 侮辱类的条件概率数组
    p1Vec - 非侮辱类的条件概率数组
    pClass1 - 文档属于侮辱类的概率
returns:

"""

def classifyNB(vec2Classify, p0Vec, p1Vec, pClass1):
    #对数相加，相当与计算两个向量相乘的结果
    #这里不太懂，感觉和公式对不上
    p1 = sum(vec2Classify * p1Vec) + log(pClass1)
    p0 = sum(vec2Classify * p0Vec) + log(1.0 - pClass1)
    if p1 > p0:
        return 1
    else: 
        return 0

"""
function:
    处理文本
parameter:
　  bigString - 文本
return:
    tok - 文本处理后所得词汇向量
"""
def textParse(bigString):
    listOfTokens = re.split(rb'\w*', bigString)
    return [tok.lower() for tok in listOfTokens if len(tok) > 2]

"""
function:
    垃圾邮件过滤
parameter:
    无
returns:
    无
"""

def spamTest():
    docList = [] #数据集
    classList = [] #标签集
    fullText = [] #???
    for i in range(1,26):
        wordList = textParse(open('spam/%d.txt' % i, 'rb').read())
        docList.append(wordList)
        fullText.extend(wordList)
        classList.append(1)
        wordList = textParse(open('ham/%d.txt' % i, 'rb').read())
        docList.append(wordList)
        fullText.extend(wordList)
        classList.append(0)

    vocabList = createVocabList(docList)#创建词汇表
    trainingSet = list(range(50))#训练集
    testSet = []#测试集
    for i in range(10):#选取10个测试集，并从训练集中删除
        randIndex = int(random.uniform(0,len(trainingSet)))
        testSet.append(trainingSet[randIndex])
        del(trainingSet[randIndex])#???
    trainMat = []#数字化的训练集
    trainClasses = []#标签集
    for docIndex in trainingSet:#遍历训练集,计算数据
        trainMat.append(setOfWords2Vec(vocabList, docList[docIndex]))
        trainClasses.append(classList[docIndex])
    p0V, p1V, pSpam = trainNB0(array(trainMat), array(trainClasses))
    
    errorCount = 0
    for docIndex in testSet:#测试集计算错误率
        wordVector= setOfWords2Vec(vocabList, docList[docIndex])
        if classifyNB(array(wordVector), p0V, p1V, pSpam) != \
      classList[docIndex]:
          errorCount += 1
    print("the error rate is: ", float(errorCount)/len(testSet))

if __name__ == "__main__":
    spamTest()

最后一个不写了