一种新的关联规则挖掘算法研究  

New mining algorithm study on association rules

作　　者：韦玉科[1] 汪仁煌[1] 李江平[1] 陈群[2] 

机构地区：[1]广东工业大学自动化学院 [2]广州中医药大学基础医学院,广州510405

出　　处：《计算机应用研究》2008年第10期2962-2964,2969,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(30472122)

摘　　要：通过分析数据关联的特点和已有的关联规则挖掘算法,在定量描述的准确性和算法高效性方面作了进一步研究,提出了更准确的支持度和置信度定量描述方法和关联关系强弱的定量描述方法。同时,改进了FP-growth挖掘算法,并应用于中医舌诊临床病例数据库挖掘实验中,可成功准确地提取中医舌诊诊断规则。测试结果表明该算法速度快、准确度高。By analyzed the characteristics of data association and the mining algorithm of association rules that have been put forward, advanced research have been made in every field of accuracy of quantitative description and high efficiency of algorithm. The article put forward a more efficiency method of quantitative description about support degree and confidence degree, and a method of quantitative description about strength of association relationship. Meanwhile the article improved FPgrowth mining algorithm and applied it to mining experiment on clinical diseases database in the traditional Chinese medicine （TCM） tongue diagnosis. The algorithm could accurately draw diagnosis rule from （TCM） tongue diagnosis successfully. The application result in make clear that, the algorithm can be faster and accurate more.

关 键 词：数据挖掘 关联规则 频繁模式增长算法 频繁模式树 中医诊断 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]