基于KNN的newsgroup 18828文本分类器的Python实现

还是同前一篇作为学习入门。

1. KNN算法描述：

step1： 文本向量化表示，计算特征词的TF-IDF值

step2： 新文本到达后，根据特征词确定文本的向量

step3 : 在训练文本集中选出与新文本向量最相近的k个文本向量，相似度度量采用“余弦相似度”，根据实验测试的结果调整k值，此次选择20

step4： 在新文本的k个邻居中，依次计算每类的权重，

step5： 比较类的权重，将新文本放到权重最大的那个类中

2. 文档TF-IDF计算和向量化表示

# -*- coding: utf-8 -*-
import time
from os import listdir
from math import log
from numpy import *
from numpy import linalg
from operator import itemgetter

###################################################
## 计算所有单词的IDF值
###################################################
def computeIDF():
    fileDir = 'processedSampleOnlySpecial_2'
    wordDocMap = {}  # <word, set(docM,...,docN)>
    IDFPerWordMap = {}  # <word, IDF值>
    countDoc = 0.0
    cateList = listdir(fileDir)
    for i in range(len(cateList)):
        sampleDir = fileDir + '/' + cateList[i]
        sampleList = listdir(sampleDir)
        for j in range(len(sampleList)):
            sample = sampleDir + '/' + sampleList[j]
            for line in open(sample).readlines():
                word = line.strip('\n')
                if word in wordDocMap.keys():
                    wordDocMap[word].add(sampleList[j]) # set结构保存单词word出现过的文档
                else:
                    wordDocMap.setdefault(word,set())
                    wordDocMap[word].add(sampleList[j])
        print 'just finished %d round ' % i

    for word in wordDocMap.keys():
        countDoc = len(wordDocMap[word]) # 统计set中的文档个数
        IDF = log(20000/countDoc)/log(10)
        IDFPerWordMap[word] = IDF

    return IDFPerWordMap

###################################################
## 将IDF值写入文件保存
###################################################
def main():
    start=time.clock()
    IDFPerWordMap = computeIDF()
    end=time.clock()
    print 'runtime: ' + str(end-start)
    fw = open('IDFPerWord','w')
    for word, IDF in IDFPerWordMap.items():
        fw.write('%s %.6f\n' % (word,IDF))
    fw.close()

########################################################
## 生成训练集和测试集的文档向量，向量形式<cate, doc, (word1, tdidf1), (word2, tdidf2),...> 存入文件
## @param indexOfSample 迭代的序号
## @param trainSamplePercent 训练集合和测试集合划分百分比
########################################################
def computeTFMultiIDF(indexOfSample, trainSamplePercent):
    IDFPerWord = {} # <word, IDF值> 从文件中读入后的数据保存在此字典结构中
    for line in open('IDFPerWord').readlines():
        (word, IDF) = line.strip('\n').split(' ')
        IDFPerWord[word] = IDF        

    fileDir = 'processedSampleOnlySpecial_2'
    trainFileDir = "docVector/" + 'wordTFIDFMapTrainSample' + str(indexOfSample)
    testFileDir = "docVector/" + 'wordTFIDFMapTestSample' + str(indexOfSample)

    tsTrainWriter = open(trainFileDir, 'w')
    tsTestWriter = open(testFileDir, 'w')

    cateList = listdir(fileDir)
    for i in range(len(cateList)):
        sampleDir = fileDir + '/' + cateList[i]
        sampleList = listdir(sampleDir)

        testBeginIndex = indexOfSample * ( len(sampleList) * (1-trainSamplePercent) )
        testEndIndex = (indexOfSample+1) * ( len(sampleList) * (1-trainSamplePercent) )

        for j in range(len(sampleList)):
            TFPerDocMap = {} # <word, 文档doc下该word的出现次数>
            sumPerDoc = 0  # 记录文档doc下的单词总数
            sample = sampleDir + '/' + sampleList[j]
            for line in open(sample).readlines():
                sumPerDoc += 1
                word = line.strip('\n')
                TFPerDocMap[word] = TFPerDocMap.get(word, 0) + 1

            if(j >= testBeginIndex) and (j <= testEndIndex):
                tsWriter = tsTestWriter
            else:
                tsWriter = tsTrainWriter

            tsWriter.write('%s %s ' % (cateList[i], sampleList[j])) # 写入类别cate，文档doc

            for word, count in TFPerDocMap.items():
                TF = float(count)/float(sumPerDoc)
                tsWriter.write('%s %f ' % (word, TF * float(IDFPerWord[word]))) # 继续写入类别cate下文档doc下的所有单词及它的TF-IDF值

            tsWriter.write('\n')

        print 'just finished %d round ' % i

        #if i==0: break

    tsTrainWriter.close()
    tsTestWriter.close()
    tsWriter.close()

3. KNN算法的实现

def doProcess():
    trainFiles = 'docVector/wordTFIDFMapTrainSample0'
    testFiles = 'docVector/wordTFIDFMapTestSample0'
    kNNResultFile = 'docVector/KNNClassifyResult'

    trainDocWordMap = {}  # 字典<key, value> key=cate_doc, value={{word1,tfidf1}, {word2, tfidf2},...}

    for line in open(trainFiles).readlines():
        lineSplitBlock = line.strip('\n').split(' ')
        trainWordMap = {}
        m = len(lineSplitBlock)-1
        for i in range(2, m, 2):  # 在每个文档向量中提取(word, tfidf)存入字典
            trainWordMap[lineSplitBlock[i]] = lineSplitBlock[i+1]

        temp_key = lineSplitBlock[0] + '_' + lineSplitBlock[1]  # 在每个文档向量中提取类目cate，文档doc，
        trainDocWordMap[temp_key] = trainWordMap 

    testDocWordMap = {}

    for line in open(testFiles).readlines():
        lineSplitBlock = line.strip('\n').split(' ')
        testWordMap = {}
        m = len(lineSplitBlock)-1
        for i in range(2, m, 2):
            testWordMap[lineSplitBlock[i]] = lineSplitBlock[i+1]

        temp_key = lineSplitBlock[0] + '_' + lineSplitBlock[1]
        testDocWordMap[temp_key] = testWordMap #<类_文件名，<word, TFIDF>>

    #遍历每一个测试样例计算与所有训练样本的距离，做分类
    count = 0
    rightCount = 0
    KNNResultWriter = open(kNNResultFile,'w')
    for item in testDocWordMap.items():
        classifyResult = KNNComputeCate(item[0], item[1], trainDocWordMap)  # 调用KNNComputeCate做分类

        count += 1
        print 'this is %d round' % count

        classifyRight = item[0].split('_')[0]
        KNNResultWriter.write('%s %s\n' % (classifyRight,classifyResult))
        if classifyRight == classifyResult:
            rightCount += 1
        print '%s %s rightCount:%d' % (classifyRight,classifyResult,rightCount)

    accuracy = float(rightCount)/float(count)
    print 'rightCount : %d , count : %d , accuracy : %.6f' % (rightCount,count,accuracy)
    return accuracy

#########################################################
## @param cate_Doc 测试集<类别_文档>
## @param testDic 测试集{{word, TFIDF}}
## @param trainMap 训练集<类_文件名，<word, TFIDF>>
## @return sortedCateSimMap[0][0] 返回与测试文档向量距离和最小的类
#########################################################
def KNNComputeCate(cate_Doc, testDic, trainMap):
    simMap = {} #<类目_文件名,距离> 后面需要将该HashMap按照value排序
    for item in trainMap.items():
        similarity = computeSim(testDic,item[1])  # 调用computeSim()
        simMap[item[0]] = similarity

    sortedSimMap = sorted(simMap.iteritems(), key=itemgetter(1), reverse=True) #<类目_文件名,距离> 按照value排序

    k = 20
    cateSimMap = {} #<类，距离和>
    for i in range(k):
        cate = sortedSimMap[i][0].split('_')[0]
        cateSimMap[cate] = cateSimMap.get(cate,0) + sortedSimMap[i][1]

    sortedCateSimMap = sorted(cateSimMap.iteritems(),key=itemgetter(1),reverse=True)

    return sortedCateSimMap[0][0]   

#################################################
## @param testDic 一维测试文档向量<<word, tfidf>>
## @param trainDic 一维训练文档向量<<word, tfidf
## @return 返回余弦相似度
def computeSim(testDic, trainDic):
    testList = []  # 测试向量与训练向量共有的词在测试向量中的tfidf值
    trainList = []  # # 测试向量与训练向量共有的词在训练向量中的tfidf值

    for word, weight in testDic.items():
        if trainDic.has_key(word):
            testList.append(float(weight)) # float()将字符型数据转换成数值型数据，参与下面运算
            trainList.append(float(trainDic[word]))

    testVect = mat(testList)  # 列表转矩阵，便于下面向量相乘运算和使用Numpy模块的范式函数计算
    trainVect = mat(trainList)
    num = float(testVect * trainVect.T)
    denom = linalg.norm(testVect) * linalg.norm(trainVect)
    #print 'denom:%f' % denom
    return float(num)/(1.0+float(denom))

输出结果：

运行时遇到几种语法错误：

Error1：

split(' ')按空格分割后最后一位是空串，不检查分割后数组的最后一位很难发现，以致产生越界

Error2：

因为导入数据用字符串运算strip切分后，返回的都是字符型数据，而string不能计算，需int(string) float(string）转换后参与计算，错误如下：

Error3：

sorted()对字典中每对<key, value>数据进行排序，返回的是包含tuple(key, value)的列表，之前不了解这一点，出现了index的错误：

sorted()的返回形式：