FP-growth算法应用实例（基于python）
我把关于FP-Growth的算法原理，python实现代码，以及代码解读放在了另外一篇文章：有兴趣可以看看。
这篇文章给出该算法的一个很酷的实例应用。我们将用到一个叫 kosarak.dat 的数据集，可以从这里下载。这份数据集包含将近100万条记录，对于展示FP-Growth算法的速度十分有效。该文件的每一行包含某个用户浏览过的新闻报道。用户和报道被编码成整数。
为了看起来方便，还是先放一下python的实现代码：
#FP-Growth实现代码
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):
self.count += numOccur
def disp(self, ind=1):
print ' '*ind, self.name, ' ', self.count
for child in self.children.values():
child.disp(ind+1)
def updateHeader(nodeToTest, targetNode):
while nodeToTest.nodeLink != None:
nodeToTest = nodeToTest.nodeLink
nodeToTest.nodeLink = targetNode
def updateFPtree(items, inTree, headerTable, count):
if items[0] in inTree.children:
# 判断items的第一个结点是否已作为子结点
inTree.children[items[0]].inc(count)
else:
# 创建新的分支
inTree.children[items[0]] = treeNode(items[0], count, inTree)
# 更新相应频繁项集的链表，往后添加
if headerTable[items[0]][1] == None:
headerTable[items[0]][1] = inTree.children[items[0]]
else:
updateHeader(headerTable[items[0]][1], inTree.children[items[0]])
# 递归
if len(items) > 1:
updateFPtree(items[1::], inTree.children[items[0]], headerTable, count)
def createFPtree(dataSet, minSup=1):
headerTable = {}
for trans in dataSet:
for item in trans:
headerTable[item] = headerTable.get(item, 0) + dataSet[trans]
for k in headerTable.keys():
if headerTable[k] < minSup:
del(headerTable[k]) # 删除不满足最小支持度的元素
freqItemSet = set(headerTable.keys()) # 满足最小支持度的频繁项集
if len(freqItemSet) == 0:
return None, None
for k in headerTable:
headerTable[k] = [headerTable[k], None] # element: [count, node]
retTree = treeNode('Null Set', 1, None)
for tranSet, count in dataSet.items():
# dataSet：[element, count]
localD = {}
for item in tranSet:
if item in freqItemSet: # 过滤，只取该样本中满足最小支持度的频繁项
localD[item] = headerTable[item][0] # element : count
if len(localD) > 0:
# 根据全局频数从大到小对单样本排序
orderedItem = [v[0] for v in sorted(localD.items(), key=lambda p:p[1], reverse=True)]
# 用过滤且排序后的样本更新树
updateFPtree(orderedItem, retTree, headerTable, count)
return retTree, headerTable
def loadSimpDat():
simDat = [['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 simDat
# 构造成 element : count 的形式
def createInitSet(dataSet):
retDict={}
for trans in dataSet:
key = frozenset(trans)
if retDict.has_key(key):
retDict[frozenset(trans)] += 1
else:
retDict[frozenset(trans)] = 1
return retDict
# 数据集
def loadSimpDat():
simDat = [['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 simDat
# 构造成 element : count 的形式
def createInitSet(dataSet):
retDict={}
for trans in dataSet:
key = frozenset(trans)
if retDict.has_key(key):
retDict[frozenset(trans)] += 1
else:
retDict[frozenset(trans)] = 1
return retDict
# 递归回溯
def ascendFPtree(leafNode, prefixPath):
if leafNode.parent != None:
prefixPath.append(leafNode.name)
ascendFPtree(leafNode.parent, prefixPath)
# 条件模式基
def findPrefixPath(basePat, myHeaderTab):
treeNode = myHeaderTab[basePat][1] # basePat在FP树中的第一个结点
condPats = {}
while treeNode != None:
prefixPath = []
ascendFPtree(treeNode, prefixPath) # prefixPath是倒过来的，从treeNode开始到根
if len(prefixPath) > 1:
condPats[frozenset(prefixPath[1:])] = treeNode.count # 关联treeNode的计数
treeNode = treeNode.nodeLink # 下一个basePat结点
return condPats
def mineFPtree(inTree, headerTable, minSup, preFix, freqItemList):
# 最开始的频繁项集是headerTable中的各元素
bigL = [v[0] for v in sorted(headerTable.items(), key=lambda p:p[1])] # 根据频繁项的总频次排序
for basePat in bigL: # 对每个频繁项
newFreqSet = preFix.copy()
newFreqSet.add(basePat)
freqItemList.append(newFreqSet)
condPattBases = findPrefixPath(basePat, headerTable) # 当前频繁项集的条件模式基
myCondTree, myHead = createFPtree(condPattBases, minSup) # 构造当前频繁项的条件FP树
if myHead != None:
# print 'conditional tree for: ', newFreqSet
# myCondTree.disp(1)
mineFPtree(myCondTree, myHead, minSup, newFreqSet, freqItemList) # 递归挖掘条件FP树运行：
#从新闻网站点击流中挖掘
parsedDat=[line.split() for line in open('kosarak.dat').readlines()]
initSet=createInitSet(parsedDat)
myFPtree,myHeaderTab=createFPtree(initSet,100000)
myFreqList=[]
mineFPtree(myFPtree,myHeaderTab,100000,set([]),myFreqList)
print "myFreqList\'s length: %s" % len(myFreqList)
for item in myFreqList:
print item运行结果：
整个运行过程仅用了十几秒。
这个运行结果表明：有9个新闻报道或报道集合曾经被10万或者更多的人浏览过。这些报道或报道集合存储在变量myFreqList中。