基于Apriori关联规则的急性胰腺炎相关因素分析-中国数字医学.PDF

数据库技术与应用 Database Technique and Application 基于Apriori关联规则的 急性胰腺炎相关因素分析 ①② ① ③ ④ 陆维嘉 严壮志 倪润洲 姚登福 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。 关键词 急性胰腺炎 胰酶 数据挖掘 关联规则 Apriori算法 Doi:10.3969/j.issn.1673-7571.2013.07.021 Analysis of Acute Pancreatitis Related Factors Based on Apriori Association Rules / LU Wei-jia, YAN Zhuang-zhi, NI Run-zhou, et al//China Digital Medicine.-2013 8(7): 70 to 73 Abstract 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 formed, 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 fo