基于Apriori关联规则的急性胰腺炎相关因素分析  

Analysis of Acute Pancreatitis Related Factors Based on Apriori Association Rules

作　　者：陆维嘉[1,2] 严壮志[1] 倪润洲[3] 姚登福[4] 

机构地区：[1]上海大学通信与信息工程学院,上海市延长路149号200072 [2]南通大学附属医院信息科,江苏省南通市西寺路20号226001 [3]南通大学附属医院消化内科,江苏省南通市西寺路20号226001 [4]南通大学附属医院科技处,江苏省南通市西寺路20号226001

出　　处：《中国数字医学》2013年第7期70-73,共4页China Digital Medicine

基　　金：上海市科委"科技创新行动计划"资助项目(编号:12DZ19408003)~~

摘　　要：目的:进行急性胰腺炎相关因素研究,通过数据挖掘形成关联规则。方法:急性胰腺炎相关资料来源于南通大学附属医院电子病历系统,收集该院2010-2012年的急性胰腺炎住院病例数据。内容包括与急性胰腺炎相关的15项临床指标。首先将各项指标数据离散化为计数资料,将源数据库映射为挖掘数据库;然后设置最小支持度为0.1,最小置信度为0.8,利用Apriori算法进行关联规则挖掘,最后由临床消化科和检验科专家解释及检验所产生的规则。结果:共形成247个强关联规则,这些强关联规则中蕴含着急性胰腺炎发病与血清酶系列及血常规指标、性别、年龄等因素之间的关联关系。结论:通过数据挖掘方法挖掘目前医院中大量医疗病例,从而发现其中蕴含的新信息、新知识,为医院大量病案信息提供了新的研究方法,具有一定的实用价值。Objective： To study the relevant factor of Acute Pancreatitis and to get the association rules by data mining. Methods： The cases with Acute Pancreatiti were obtained from 2010 to 2012 in the affilicated hospital of NanTong University. 15 indexes about Acute Pancreatitis were observed. First each index was discredited into count data. And source database was mapped to mining database. Then set the minimum support of 0.1, minimum confidence of 0.8, using Apriori algorithm for mining association rules, and finally by the Digesting internal medicine specialists and Clinical laboratory experts to explain and test the rules generated. Results： 247strong association rules were foruled, containing the relationships of onset of Acute Pancreatitis with the sex, age, function of pancreatic enzyme, and parameters of blood. Conclusion： The method of this study could new information by the data mining. It was valuable for supplying a new approach tu assess the huge medical information.

关 键 词：急性胰腺炎 胰酶 数据挖掘 关联规则 APRIORI算法 

分 类 号：R576[医药卫生—消化系统]