1 #生成关联规则
 2 #L: 频繁项集列表
 3 #supportData: 包含频繁项集支持数据的字典
 4 #minConf 最小置信度
 5 def generateRules(L, supportData, minConf=0.7):
 6     #包含置信度的规则列表
 7     bigRuleList = []
 8     #从频繁二项集开始遍历
 9     for i in range(1, len(L)):
10         for freqSet in L[i]:
11             H1 = [frozenset([item]) for item in freqSet]
12             if (i > 1):
13                 rulesFromConseq(freqSet, H1, supportData, bigRuleList, minConf)
14             else:
15                 calcConf(freqSet, H1, supportData, bigRuleList, minConf)
16     return bigRuleList
17
18
19 # 计算是否满足最小可信度
20 def calcConf(freqSet, H, supportData, brl, minConf=0.7):
21     prunedH = []
22     #用每个conseq作为后件
23     for conseq in H:
24         # 计算置信度
25         conf = supportData[freqSet] / supportData[freqSet - conseq]
26         if conf >= minConf:
27             print(freqSet - conseq, '-->', conseq, 'conf:', conf)
28             # 元组中的三个元素：前件、后件、置信度
29             brl.append((freqSet - conseq, conseq, conf))
30             prunedH.append(conseq)
31
32     #返回后件列表
33     return prunedH
34
35
36 # 对规则进行评估
37 def rulesFromConseq(freqSet, H, supportData, brl, minConf=0.7):
38     m = len(H[0])
39     if (len(freqSet) > (m + 1)):
40         Hmp1 = aprioriGen(H, m + 1)
41        # print(1,H, Hmp1)
42         Hmp1 = calcConf(freqSet, Hmp1, supportData, brl, minConf)
43         if (len(Hmp1) > 0):
44             rulesFromConseq(freqSet, Hmp1, supportData, brl, minConf)