关联规则挖掘在电子病历分析中的应用研究  

Application and research of association rule on data mining for analysis of CPR

作　　者：刘秀娜[1] 柏建普[1] 

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《内蒙古科技大学学报》2010年第4期360-363,共4页Journal of Inner Mongolia University of Science and Technology

摘　　要：针对甲状腺电子病历数据量大、更新速度快的特点,提出了一种挖掘有效关联规则的技术.该技术运用区间归并法与特征区间法相结合的离散化方法对病历中的数据进行预处理;在规则生成的核心算法中,提出了优化的增量更新FUP算法,算法通过对新旧数据库设定不同的支持度得到病历各属性间的关联规则.实验验证了改进算法的有效性,挖掘结果对了解疾病的诊断、治疗、发展规律有重要价值,对医学研究有重要意义.CPR,the abbreviation of computer-based patient record has the feature that they have large volumes of data as well as update speed.Based on these,a new technology was proposed to mine valid association rules.The technology uses the interval merging method combined with the characteristic interval method to discrete data in the process of data pretreatment.An optimized incremental update FUP algorithm in the core algorithm to generate association rules was presented,obtaining the relative rules among various properties by setting different support between the old and new database.The effectiveness of the improved algorithm was tested by experiment.The mining results may play an role in understanding the diagnosis,treatment and development of diseases.

关 键 词：数据挖掘 关联规则 电子病历 FUP算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]