中文文本聚类

使用python实现中文文本聚类，利用kmeans算法，包含jiba分词方法等

‘‘‘
Created on Mar 24 2011
Ch 11 code
@author: Peter
‘‘‘
from numpy import *

def loadDataSet（）:
    return [
         [12]
        [2345]
        [1346]
        [2134]
        [2136]
    ]

def createC1（dataSet）:
    C1 = []
    for transaction in dataSet:
        for item in transaction:
            if not [item] in C1:
                C1.append（[item]）
                
    C1.sort（）
    return map（frozenset C1）#use frozen set so we
                            #can use it as a key in a dict    

def scanD（D Ck minSupport）:
    ssCnt = {}
    for tid in D:
        for can in Ck:
            if can.issubset（tid）:
                if not ssCnt.has_key（can）: ssCnt[can]=1
                else: ssCnt[can] += 1
    numItems = float（len（D））
    retList = []
    supportData = {}
    for key in ssCnt:
        support = ssCnt[key]/numItems
        if support >= minSupport:
            retList.insert（0key）
        supportData[key] = support
    return retList supportData

def aprioriGen（Lk k）: #creates Ck
    retList = []
    lenLk = len（Lk）
    for i in range（lenLk）:
        for j in range（i+1 lenLk）: 
            L1 = list（Lk[i]）[:k-2]; 
            L2 = list（Lk[j]）[:k-2]
            L1.sort（）; 
            L2.sort（）
            print L1L2
            if L1==L2: #if first k-2 elements are equal
                retList.append（Lk[i] | Lk[j]） #set union
    return retList

def apriori（dataSet minSupport = 0.5）:
    C1 = createC1（dataSet）
    D = map（set dataSet）
    L1 supportData = scanD（D C1 minSupport）
    L = [L1]
    k = 2
    while （len（L[k-2]） > 0）:
        Ck = aprioriGen（L[k-2] k）
        Lk supK = scanD（D Ck minSupport）#scan DB to get Lk
        supportData.update（supK）
        L.append（Lk）
        k += 1
    return L supportData

def generateRules（L supportData minConf=0.7）:  #supportData is a dict coming from scanD
    bigRuleList = []
    for i in range（1 len（L））:#only get the sets with two or more items
        for freqSet in L[i]:
            H1 = [frozenset（[item]） for item in freqSet]
            if （i > 1）:
                rulesFromConseq（freqSet H1 supportData bigRuleList minConf）
            else:
                calcConf（freqSet H1 supportData bigRuleList minConf）
    return bigRuleList         

def calcConf（freqSet H supportData brl minConf=0.7）:
    prunedH = [] #create new list to return
    for conseq in H:
        conf = supportData[freqSet]/supportData[freqSet-conseq] #calc confidence
        if conf >= minConf: 
            print freqSet-conseq‘-->‘conseq‘conf:‘conf
            brl.append（（freqSet-conseq conseq conf））
            prunedH.append（conseq）
    return prunedH

def rulesFromConseq（freqSet H supportData brl minConf=0.7）:
    m = len（H[0]）
    if （len（freqSet） > （m + 1））: #try further merging
        Hmp1 = aprioriGen（H m+1）#create Hm+1 new candidates
        Hmp1 

 属性            大小     日期    时间   名称
----------- ---------  ---------- -----  ----
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\
     文件        6012  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\apriori.py
     文件        5189  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\apriori.pyc
     文件       38906  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\bills20DataSet.txt
     文件      137426  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\lawAssnRules.txt
     文件        1806  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\meaning20.txt
     文件      570408  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\mushroom.dat
     文件        5585  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\recent100bills.txt
     文件        1050  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\Apriori\recent20bills.txt
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\FP-growth\
     文件        6615  2015-10-11 04:20  chinese_text_cluster-master\Association_Analysis\FP-growth\fpGrowth.py
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Classification\
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\EXTRAS\
     文件         522  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\EXTRAS\README.txt
     文件         784  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\EXTRAS\simpleDataPlot.py
     文件        5548  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\adaboost.py
     文件        4719  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\adaboost.pyc
     文件       13614  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\horseColicTest2.txt
     文件       60778  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\horseColicTraining2.txt
     文件        3462  2015-10-11 04:20  chinese_text_cluster-master\Classification\AdaBoost\old_adaboost.py
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\
     目录           0  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\EXTRAS\
     文件         522  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\EXTRAS\README.txt
     文件         961  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\EXTRAS\create2Normal.py
     文件         456  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\EXTRAS\monoDemo.py
     文件        7247  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\bayes.py
     文件        6957  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\bayes.pyc
     文件       15141  2015-10-11 04:20  chinese_text_cluster-master\Classification\Bayes\email.zip
............此处省略3084个文件信息