一种快速的自顶向下挖掘算法  

A FAST TOP-DOWN MINING ALGORITHM

作　　者：胡宁[1] 

机构地区：[1]重庆三峡学院,重庆404000

出　　处：《计算机应用与软件》2014年第5期272-275,共4页Computer Applications and Software

基　　金：重庆市自然科学基金项目(CSTC2010BB7316);重庆三峡学院科学技术研究项目(13ZD20)

摘　　要：提出一种基于按序产生子集的关联规则挖掘算法,适合于挖掘长频繁项目集;算法基于自顶向下搜索策略,采用按序产生子集的方法生成频繁候选项,在挖掘过程中减少了冗余候选项和重复计算,提高了挖掘效率。通过模拟实验数据与现有经典挖掘算法进行性能测试比较,结果表明当它挖掘长频繁项目集时不仅快速而且有效。The article presents an association rules mining algorithm based on generating subsets in sequence. It is especially suitable for mining long frequent itemsets. The algorithm is founded on top-down search strategy,utilizing a ＂ generating subsets in sequence＂ method to generate frequent candidate items. During the mining process the number of redundant candidate items and repeated computations is reduced and the mining efficiency is improved. By performance testing comparison with simulation experimental data against present classical mining algorithm,it is proven that,when mining long frequent itemsets,it is fast and effective.

关 键 词：长频繁项目集 自顶向下 关联规则 数据挖掘 算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]