一种面向中医药领域的二维最大频繁项集挖掘算法  

Two-dimensional Maximal Frequent Itemset Mining Algorithm in Traditional Chinese Medicine Field

作　　者：肖文[1] 鞠时光[1] 朱金伟[1] 辛燕[1] 刘志锋[1] 

机构地区：[1]江苏大学计算机学院,江苏镇江212013

出　　处：《小型微型计算机系统》2007年第12期2193-2198,共6页Journal of Chinese Computer Systems

基　　金：江苏省自然科学基金(BK2006073)资助;江苏大学青年基金项目(1241170006)资助

摘　　要：在中医药领域挖掘药组频繁项集时发现,尽管有些项集的支持度比人们需要的频繁项集的支持度高很多,但这些项集并不是人们感兴趣的,即过分频繁反而变得平凡.本文引入支持度区间的概念,提出了适合中药数据挖掘的二维TCM-FP森林结构及其建树算法.在针对疾病症状的中药药组挖掘过程中,采用优化的搜索策略开发了基于支持度区间的TCMA维间最大频繁项集挖掘算法.这种算法既缩小了挖掘的范围又提高了规则的意义,并且具有较高的执行效率.When mining medicine group in TCM （traditional Chinese medicine） field, it shows that although some itemsets＇ support rating is higher than those frequent itemsets need, these itemsets are not interesting, namely, too frequent makes them ordinary. In this paper, the concept of support rating interval is brought in and the structure of two-dimensional TCM-FP forest and the algorithm of tree building are proposed for the TCM data mining. While mining TCM medicine groups for disease symptoms, optimized search strategy is adopted and the TCMA mining algorithm of intra-dimensional maximal frequent itemsets based on support rating interval is developed. This algorithm both reduces the mining scope and improves the significance of the rules, and has higher execute efficiency than FP-growth.

关 键 词：数据挖掘 中药 维间关联规则 最大频繁项集 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]