Python实现FP-growth【链表】【递归】

糖逗

参考书籍：《机器学习实战》

import numpy as np

class treeNode:
    def __init__(self, nameValue, numOccur, parentNode):
        self.name = nameValue
        self.count = numOccur
        self.nodeLink = None
        self.parent = parentNode
        self.children = {}
        
    def inc(self, numOccur):#增加当前节点的值
        self.count += numOccur
    
    def disp(self, ind = 1):#打印树，ind设置成0也可以吗？？？？？
        print('  '*ind, self.name, ' ', self.count)
        for child in self.children.values():
            child.disp(ind + 1)#递归
 

def createTree(dataSet, minSup = 1):#扫描头指针表，删掉哪些出现次数少于minSup的项
    headerTable = {}
    for trans in dataSet:
        for item in trans:
            #https://www.runoob.com/python/att-dictionary-get.html
            headerTable[item] = headerTable.get(item, 0) + dataSet[trans]#统计每个字符出现的次数
    for k in list(headerTable.keys()):  
        if headerTable[k] < minSup: 
            del(headerTable[k])
    freqItemSet = set(headerTable.keys())#通过minSup筛选的频繁集
    if len(freqItemSet) == 0:#如果每个元素只出现一次，那就直接返回None（没有规律可循） 
        return None, None  
    for k in headerTable:
        headerTable[k] = [headerTable[k], None]#None可能存储指针
    retTree = treeNode('Null Set', 1, None)#根节点设置成空集
    for tranSet, count in dataSet.items():  
        localD = {}
        for item in tranSet:  
            if item in freqItemSet:#单独出现
                localD[item] = headerTable[item][0]
        if len(localD) > 0:
            orderedItems = [v[0] for v in sorted(localD.items(), 
                            key = lambda p: p[1], reverse = True)]#降序排列
            updateTree(orderedItems, retTree, headerTable, count)
    return retTree, headerTable 

def updateTree(items, inTree, headerTable, count):
    if items[0] in inTree.children:#如果在节点中，直接更新节点count
        inTree.children[items[0]].inc(count) 
    else:  
        inTree.children[items[0]] = treeNode(items[0], count, inTree)#如果没有则新增节点
        if headerTable[items[0]][1] == None:#当前item不属于频繁集中，更新headerTable
            headerTable[items[0]][1] = inTree.children[items[0]]
        else:
            updateHeader(headerTable[items[0]][1], inTree.children[items[0]])
    if len(items) > 1:
        updateTree(items[1:], inTree.children[items[0]], headerTable, count)
        
def updateHeader(nodeToTest, targetNode):   
    while (nodeToTest.nodeLink != None):#移动到链表的最后端 
        nodeToTest = nodeToTest.nodeLink
    nodeToTest.nodeLink = targetNode#将链表最后段关联当前端        

def ascendTree(leafNode, prefixPath):#从叶子节点递归至根节点
    if leafNode.parent != None:
        prefixPath.append(leafNode.name)
        ascendTree(leafNode.parent, prefixPath)
    
def findPrefixPath(basePat, treeNode): #headerTable中存储的是元素链表的起始指针
    condPats = {}
    while treeNode != None:
        prefixPath = []
        ascendTree(treeNode, prefixPath)
        if len(prefixPath) > 1: 
            condPats[frozenset(prefixPath[1:])] = treeNode.count
        treeNode = treeNode.nodeLink
    return condPats

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

def createInitSet(dataSet):
    retDict = {}
    for trans in dataSet:
        retDict[frozenset(trans)] = 1
    return retDict

def mineTree(inTree, headerTable, minSup, preFix, freqItemList):
    bigL = [v[0] for v in sorted(headerTable.items(), key = lambda p: p[1][0])]
    for basePat in bigL: 
        newFreqSet = preFix.copy()
        newFreqSet.add(basePat)
        freqItemList.append(newFreqSet)
        condPattBases = findPrefixPath(basePat, headerTable[basePat][1])#提取前缀路径
        myCondTree, myHead = createTree(condPattBases, minSup)#根据前缀路径构建条件FP树
        if myHead != None:        
            mineTree(myCondTree, myHead, minSup, newFreqSet, freqItemList)#从二元扩展到三元，从三元扩展到四元，以此类推递归
    return
            
if __name__ == "__main__":
    simpDat = loadSimpDat()
    initSet = createInitSet(simpDat)
    myFPtree, myHeaderTab = createTree(initSet, 3)
    #myFPtree.disp()
    #res = findPrefixPath('r', myHeaderTab['r'][1])
    freqItem = []
    mineTree(myFPtree, myHeaderTab, 3, set(), freqItem)
    print(freqItem)

糖逗

未完待续，明天接着学习。