摘要
为解决在挖掘关联规则时存在大量冗余规则以及效率不高的问题,提出了一种基于事务ID集合的带约束的关联规则挖掘算法ACARMT。该算法结合了Separate算法以及基于数据垂直分布算法的优势,先根据约束条件产生基础频繁项目集,再利用事务ID集合存储项目集信息,从而避免重复扫描数据库,提高了挖掘效率。应用该算法挖掘实际的生殖健康数据的实验表明,在数据量大到超出基于数据垂直分布算法的使用范围时,该算法仍然有效,并且其效率优于Sepa-rate算法。
To deal with the problems which there are lots of redundant rules and the efficiency is not high when extracting association rules, an algorithm of constrained association rules mining based on the set of transaction id (ACARMT) is proposed. This algorithm combines the advantages of the algorithm named Separate and the algorithm based on vertical data. Firstly, basal frequent items are generated by the constraint condition. Then the items' information is stored by set of transaction id to avoid scanning database repeatedly and improve the mining efficiency. Finally, the results of using the proposed ACARMT to mining actual reproductive health data indicate ACARMT has better efficiency than Separate, and can function well when the scale of data is too large to use the algorithm based on vertical data.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第5期1663-1667,共5页
Computer Engineering and Design
基金
国家科技支撑计划课题基金项目(2009BAH39B03)
国家自然科学基金项目(61072060)
国家高技术研究发展计划课题基金项目(2011AA100706)
高等学校博士学科点专项科研基金课题基金项目(20110005120007)
中央高校基本科研业务费专项基金项目(2012RC0205)