检测联合用药不良反应信号的数据挖掘方法  

Comparison of Data Mining Methodologies Used to Detect Drug-drug interactions

作　　者：钱轶峰[1] 罗宝章[1] 叶小飞[1] 孙亚林[1] 吴美京[1] 王海南[1,2] 杜文民[3] 贺佳[1] 

机构地区：[1]第二军医大学卫勤系卫生统计学教研室,200433 [2]国家食品药品监督管理局药品审评中心,100038 [3]上海市药品不良反应监测中心,200040

出　　处：《中国卫生统计》2010年第1期31-34,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金(30872186);上海市优秀学科带头人计划(A类)(09XD1405500);新药临床研究技术平台建设分课题:数据管理和统计分析的平台建设(2008ZX09312-007和2008ZX09312-025)

摘　　要：由药物联合应用导致的不良反应(ADR)信号有其自身的特点,因此,针对此类ADR进行信号发掘的数据挖掘算法亦同一般的ADR数据挖掘算法有所区别。本文针对用于挖掘该类信号的数据挖掘算法进行了简要的介绍,并希望通过对各种算法效能及优缺点的简单比较,从而为其在实际数据库中的推广使用提供参考。此外,本文还对现有方法的扩展使用及有望在该领域发挥更大作用的其他数据挖掘算法进行了适度的探讨,以期对该领域的后续研究有所启示。There are certain characteristics lies in the signals of the adverse drug reactions（ADR）caused by drug-drug interactions,which accounts for the difference of the data mining methodologies compared with the generally used methods in the ADR detection field.In this paper,we made a brief introduction of the currently used methods in this field,the efficacy as well as the merits and drawbacks of the methods were also compared to support its further application in different databases.In addition,a deliberate discussion for the extension of the methods and some other promising tools that have not been introduced into this area was added at the end of the paper to aid the future work as much as possible.

关 键 词：联合用药 不良反应 信号挖掘 

分 类 号：R195[医药卫生—卫生统计学]