Apriori算法如何进行关联分析

def createC1(dataSet):    C1 = []    for transaction in dataSet:        for item in transaction:            if not [item] in C1:                C1.append([item])                    C1.sort()    return list(map(frozenset, C1))

def scanD(D, Ck, minSupport):    ssCnt = {}    for tid in D:        for can in Ck:            if can.issubset(tid):                if can not in ssCnt:                    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(0,key)        supportData[key] = support    return retList, supportData#返回符合支持度的子项，以及所有项目计算的支持度

#构建多项的频繁项目集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()            if L1==L2: #if first k-2 elements are equal                retList.append(Lk[i] | Lk[j]) #set union    return retList#整体Apriori函数代码,得到所有频繁项集和符合最小支持度要求的集合def apriori(dataSet, minSupport = 0.5):    C1 = createC1(dataSet)    D = list(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)#update() 函数把字典dict2的键/值对更新到dict里        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]:            print(L[i],freqSet)            H1 = [frozenset([item]) for item in freqSet]            print("i",i,"H1",H1)            if (i > 1):                rulesFromConseq(freqSet, H1, supportData, bigRuleList, minConf)            else:                calcConf(freqSet, H1, supportData, bigRuleList, minConf)    return bigRuleListdef 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 prunedHdef rulesFromConseq(freqSet, H, supportData, brl, minConf=0.7):    print("freqSet:",freqSet)    m = len(H[0])    if (len(freqSet) > (m + 1)): #try further merging        Hmp1 = aprioriGen(H, m+1)#create Hm+1 new candidates        print("Hmp1:",Hmp1)        Hmp1 = calcConf(freqSet, Hmp1, supportData, brl, minConf)        print("Hmp1:",Hmp1)        if (len(Hmp1) > 1):    #need at least two sets to merge            print("len(Hmp1):",len(Hmp1))            rulesFromConseq(freqSet, Hmp1, supportData, brl, minConf)            print("__")