利用Hash树生成频繁项目集的新方法  

Generating Frequent Itemsets Using Hash Tree

作　　者：吉根林[1,2] 赵斌[2] 孙志挥[1] 

机构地区：[1]东南大学计算机系 [2]南京师范大学计算机系

出　　处：《小型微型计算机系统》2004年第10期1841-1843,共3页Journal of Chinese Computer Systems

基　　金：国家自然科学基金 ( 79970 0 92 )资助;江苏省教育厅自然科学基金 ( 2 0 0 1SXXTSJB112 )的资助

摘　　要：频繁项目集的生成是关联规则挖掘中的关键问题 .提出基于 Hash树的频繁项目集生成新方法 ,探讨了 Hash树中候选项目集的数据组织与建立算法 ,提出了利用 Hash树计算候选项目集支持数的算法 ,并用 Java语言实现了该算法 ,最后通过实验验证了利用The generation of frequent itemsets is a key problem of mining association rules. Develop ed hash tree based algorithm for generation of frequent k itemsets. The definition and the creation algorithm of hash tree were introduced. The problem of discovering frequent k itemsets can be solved by constructing a hash tree that contains a candidate set of frequent k itemsets first and then, counting support number of candidate frequent k itemsets by traveling the hash tree. The performance of hash tree based algorithm is studied andcompared with Apriori algorithm. The experiment shows that the hash tree based algorithm is valid and efficient.

关 键 词：HASH树 频繁项目集 关联规则 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]