使用FP-growth算法高效发现频繁项集

基本内容：发现事务数据中的公共模式、FP-growth算法、发现Twitter源中的共现词

fp_growth.py

'''
Created on 2018年8月3日

@author: hcl
'''
class treeNode:
    #name存放节点名称
    #count存放计数值
    #nodeLink链接相似元素
    #parent存放父节点
    #children存放子节点
    def __init__(self, nameValue, numOccur, parentNode):
        self.name = nameValue
        self.count = numOccur
        self.nodeLink = None
        self.parent = parentNode      #needs to be updated
        self.children = {} 

    def inc(self, numOccur):
        self.count += numOccur
    #展示树结构
    def disp(self, ind=1):
        #'  '*ind表示输出ind个'  '
        print('  '*ind, self.name, ' ', self.count)
        for child in self.children.values():
        #注意使用了递归的方式对不同层级的node输出不同个数的空格
            child.disp(ind+1)

#create FP-tree from dataset but don't mine
def createTree(dataSet, minSup=1): 
    headerTable = {}
    #go over dataSet twice
    #first pass counts frequency of occurance
    #遍历两次数据集，第一次过滤出满足最小支持度的项
    for trans in dataSet:
        for item in trans:
            headerTable[item] = headerTable.get(item, 0) + dataSet[trans]
    #remove items not meeting minSup
    #删除不满足最小支持度的项,这里我与书里写的不一样,因为按书里写的会报dictionary changed size during iteratation错，这样就没问题了
    headerTable = {k:v for k,v in headerTable.items() if v >= minSup}
    freqItemSet = set(headerTable.keys())
    #print 'freqItemSet: ',freqItemSet
    if len(freqItemSet) == 0: return None, None  #if no items meet min support -->get out
    for k in headerTable:
        headerTable[k] = [headerTable[k], None] #reformat headerTable to use Node link 
    #print 'headerTable: ',headerTable
    #create tree
    #创建根节点
    retTree = treeNode('Null Set', 1, None) 
    #go through dataset 2nd time
    #第二次遍历数据集，创建树
    for tranSet, count in dataSet.items():  
        #localD存储tranSet中满足最小支持度的项和对应的出现频率
        localD = {}
        for item in tranSet:  #put transaction items in order
            if item in freqItemSet:
                localD[item] = headerTable[item][0]
        if len(localD) > 0:
        #这段代码表示,localD中没项通过计数值进行逆序排序,然后将排好序的键存放到新建的列表orderedItems中
            orderedItems = [v[0] for v in sorted(localD.items(), key=lambda p: p[1], reverse=True)]
            #populate tree with ordered freq itemset
            #用排好序的频繁项集创建树
            updateTree(orderedItems, retTree, headerTable, count)
    return retTree, headerTable #return tree and header table

def updateTree(items, inTree, headerTable, count):
    #如果items[0]项在树中,那么增加树种对应节点计数值
    if items[0] in inTree.children:#check if orderedItems[0] in retTree.children
        inTree.children[items[0]].inc(count) #incrament count
    #如果items[0]不在树中,那么创建新建对应节点,创建分支,更新headerTable头指针
    else:   #add items[0] to inTree.children
        inTree.children[items[0]] = treeNode(items[0], count, inTree)
        #items[0]对应的头指针为空,将items[0]项对应的树节点放入headerTable中
        if headerTable[items[0]][1] == None: #update header table 
            headerTable[items[0]][1] = inTree.children[items[0]]
        #如果items[0]对应的头指针非空,调用updateHeader()更新头指针
        else:
            updateHeader(headerTable[items[0]][1], inTree.children[items[0]])
    #如果频繁项集长度大于一，那么递归调用updateTree(),items[1::]表示使用频繁项集中第二项之后的所有项,inTree.children[itemes[0]]表示使用子节点
    if len(items) > 1:#call updateTree() with remaining ordered items
        updateTree(items[1::], inTree.children[items[0]], headerTable, count)

#this version does not use recursion        
def updateHeader(nodeToTest, targetNode):  
    #Do not use recursion to traverse a linked list!
    #迭代指针链表，直到nodeToTest的指针为空
    while (nodeToTest.nodeLink != None):   
        nodeToTest = nodeToTest.nodeLink
    #将targetNode加入指针链表
    nodeToTest.nodeLink = targetNode

def loadSimpDat():
    simpDat = [['r', 'z', 'h', 'j', 'p'],
               ['z', 'y', 'x', 'w', 'v', 'u', 't', 's'],
               ['z'],
               ['r', 'x', 'n', 'o', 's'],
               ['y', 'r', 'x', 'z', 'q', 't', 'p'],
               ['y', 'z', 'x', 'e', 'q', 's', 't', 'm']]
    return simpDat
#将数据集转化为字典形式,注意使用的frozenset()，因为以集合作为key，必须为frozenset类型
def createInitSet(dataSet):
    retDict = {}
    for trans in dataSet:
    #将每个集合对应计数设置为1
        retDict[frozenset(trans)] = 1
    return retDict

#ascends from leaf node to root
#从叶子节点回溯到根节点
def ascendTree(leafNode, prefixPath): 
    if leafNode.parent != None:
        prefixPath.append(leafNode.name)
        ascendTree(leafNode.parent, prefixPath)

def findPrefixPath(basePat, treeNode): #treeNode comes from header table
    condPats = {}
    #通过链表遍历相似元素项,调用ascendTree获取相似元素项的前缀路径
    while treeNode != None:
        prefixPath = []
        ascendTree(treeNode, prefixPath)
        #注意,盗用ascendTree会把叶子节点自己也添加到prefixPath中
        if len(prefixPath) > 1: 
        #注意,这里使用了数组过滤,把叶子节点过滤掉，得出前缀路径，放入字典中,{prefixPath:count}
            condPats[frozenset(prefixPath[1:])] = treeNode.count
        treeNode = treeNode.nodeLink
    return condPats

def mineTree(inTree, headerTable, minSup, preFix, freqItemList):
    #(sort header table)
    #这一行与书中不一样在于p[1][0]，通过headerTable中项对应的个数来排序
    bigL = [v[0] for v in sorted(headerTable.items(), key=lambda p: p[1][0])]
    #start from bottom of header table
    #迭代满足最小支持度的频繁项,找出条件基，递归构建条件树，挖掘出频繁项集
    for basePat in bigL:  
    #注意这里使用了preFix.copy()，因为用到了递归，preFix在同一层的项中重复使用,如果直接引用会出问题
        newFreqSet = preFix.copy()
        #将当前频繁项加入新的频繁项集
        newFreqSet.add(basePat)
        #print 'finalFrequent Item: ',newFreqSet   
        #append to set
        #将新频繁项集加入频繁项集列表
        freqItemList.append(newFreqSet)
        #找出当前频繁项的前缀路径，用于构建FP条件树挖掘新的频繁项集
        condPattBases = findPrefixPath(basePat, headerTable[basePat][1])
        #print 'condPattBases :',basePat, condPattBases
        #2. construct cond FP-tree from cond. pattern base
        #使用当前频繁项的前缀路径构建FP条件树
        myCondTree, myHead = createTree(condPattBases, minSup)
        #print 'head from conditional tree: ', myHead
        #如果FP条件树存在，那么递归
        if myHead != None: #3. mine cond. FP-tree
#             print ('conditional tree for: ',newFreqSet)
            myCondTree.disp(1)            
            mineTree(myCondTree, myHead, minSup, newFreqSet, freqItemList)

if __name__ == '__main__':
#     rootNode = treeNode("pyramid", 9, None)
#     rootNode.children["eye"] = treeNode("eye",13, None)
#     rootNode.disp()
#     rootNode.children["phoenix"] = treeNode("phoenix", 3, None)
#     rootNode.disp()
    
    
#     simpDat = loadSimpDat()
# #     print (simpDat)
#     initSet = createInitSet(simpDat)
# #     print (initSet)
#     myFPtree, myHeaderTab = createTree(initSet, 3) 
#     myFPtree.disp()


    simpDat = loadSimpDat()
    initSet = createInitSet(simpDat)
    myFPtree, myHeaderTab = createTree(initSet, 3) 
    freqItems = []
    mineTree(myFPtree, myHeaderTab, 3, set([]), freqItems)
    for item in freqItems:
        print(item)

输出：

   Null Set   1
     z   3
       x   3
   Null Set   1
     z   3
   Null Set   1
     y   2
       z   2
         x   2
     z   1
       y   1
         x   1
   Null Set   1
     x   3
   Null Set   1
     z   3
{'r'}
{'y'}
{'y', 'z'}
{'y', 'x'}
{'y', 'z', 'x'}
{'t'}
{'y', 't'}
{'z', 't'}
{'x', 't'}
{'s'}
{'s', 'x'}
{'x'}
{'z', 'x'}
{'z'}