使用Apriori算法揭示脑卒中的频繁集  

A revelation about frequent itemset of stroke by using Apriori algorithm

作　　者：孙文桥[1] 石磊[1] 何健[1] SUN Wen-qiao;SHI Lei;HE Jian(Department of Information, The 210th Hospital of PLA, Dalian 116021, China)

机构地区：[1]解放军第210医院信息科,辽宁大连116021

出　　处：《中国医学装备》2018年第3期32-35,共4页China Medical Equipment

摘　　要：目的:通过大规模脑卒中初筛数据分析,研究脑卒中发病的危险因素规则模式。方法:采集国家脑卒中筛查与防控数据中心的脑卒中危险因素初筛数据,运用关联规则挖掘方法,研究大范围人群中脑卒中发病相关的危险因素,选择最小支持度为0.1%,最小置信度为10%,挖掘脑卒中发病的模式规则。结果:在初筛表列举的9个危险因素中,既往有短暂性脑缺血发作(TIA)、高血压和脑卒中家族史与脑卒中发病呈弱相关。TIA、房颤或瓣膜性心脏病、脑卒中家族史、高血压以及糖尿病是影响脑卒中发病的最主要的危险因素。当年龄>60岁时,年龄则成为影响脑卒中发病的重要危险因素之一。结论:在21个大概率导致脑卒中发病规则模式中的一些危险因素组合此前并未得到足够的重视,脑卒中发病的危险因素规则模式为临床医学提供了新的有价值的诊断依据。Objective:To study the rule pattern of risk factor of stroke through data analysis of preliminary screening of large-scale stroke.Methods:The preliminary screening data of risk factors in stroke were gathered from the National Center for Stroke Screening and Prevention and Control.And the association rules mining was adopted in the research.The relevant risk factors of stroke in large-scale crowd were researched,and0.1%was chose as minimum support and10%were used as minimum confidence coefficient.The rule mode of stroke was mined by using above conditions.Results:In the nine risk factors of the table of preliminary screening,there was a weak correlation between"transient ischemic attack(TIA)in past","hypertension","family history of stroke"and stroke,respectively.And TIA,atrial fibrillation or valvular heart disease,family history of stroke,hypertension and diabetes were the most important risk factors.At the same time,the study found that age would become an important risk factor for stroke when patient was older than60.Conclusions:In the21presumptive rule patterns with big probability,some of them has not received sufficient attention in past.And the rule pattern of risk factor of stroke provides a new diagnostic basis with value.

关 键 词：脑卒中 危险因素规则模式 关联规则挖掘 APRIORI算法 

分 类 号：R-058[医药卫生]