摘要：

基于支持度和置信度模型的关联规则剪枝算法会挖掘出很多无趣规则。针对该问题，提出一种正相关性指导下的关联规则剪枝算法。利用全置信度和提升度构造一个正相关性评价函数，以此对频繁项集进行剪枝。实验结果表明，该算法能减少无趣关联规则数量，提升挖掘结果质量，缩短挖掘时间。

关键词: 数据挖掘, 关联规则, 兴趣度, 正相关, 剪枝

Abstract:

There are some uninteresting rules in the large of the rules excavated by the classic association rule pruning algorithm based on the support and confidence model. This paper presents a pruning algorithm based on the all confidence degree and lift degree, and it constructs a positive correlation evaluation function to pruning the frequent itemsets. Experimental results show that the algorithm can effectively reduce the number of the uninteresting correlation rule, and promote the excavation quality, and shorten excavation time.

Key words: data mining, association rule, interestingness, positive correlation, pruning

中图分类号: 

• TP301.6