关联规则挖掘在2型糖尿病并发症中的应用研究-第三军医大学学报.DOCVIP

PAGE PAGE 1 关联规则挖掘在2型糖尿病并发症中的应用研究 程远 曾照芳* （重庆医科大学检验系，临床检验诊断学教育部重点实验室，重庆 400016） [摘要] 目的：用关联规则挖掘方法研究2型糖尿病与其三种主要并发症（高脂血症、高血压、冠心病）之间的关系，为2型糖尿病并发症的早期预防提供措施。方法：利用糖尿病随访数据集中涉及到的病史情况、诊断指标、体检指标和检验参数等方面的2104条记录，选用关联规则挖掘中的Apriori算法进行分析。结果：数据分析矩阵和Apriori算法建模所得出的结果一致，即2型糖尿病与并发症的关系按照密切程度排序为: 高血压冠心病高脂血症。结论：对2型糖尿病患者可以有针对性的采取某些并发症的检查和预防措施，为2型糖尿病并发症的早期预防提供参考。 [关键词] 关联规则；数据挖掘；糖尿病；并发症 The Study of the Application for Association Rules Mining in Complications of Type 2 Diabetes Cheng yuan Zeng Zhaofang* (Key Laboratory of Laboratory Medical Diagnostics ,Ministry of Education ,Chongqing Medical University, Chongqing 400016) Abstract: Objective: To study the relationship between type 2 diabetes and its three main complications hyperlipidemia, hypertension, coronary heart disease by methods of Association Rules Mining and provide measures for early preventing these complications. Mehtods: The Apriori algorithm in Association Rules Mining was used to analyze 2,104 records collected from randomly visited diabetes patients’data that is related to medical history, diagnostic index, physical examination index and examination parameters, etc; Results: The results from Data Analyzing Matrix and Apriori algorithm modeling were the same. The order of complocations according to the close degree of relationship with type 2 diabetes were: HypertensionCoronary Heart Disease Hyperlipidemia. Conclusion: Examinations and precautions against the complications should be offered to the type 2 diabetes patients, and this study provides the reference of early prevention of the complications of type 2 diabetes. Key words: Association rules；Data mining；Diabetes；Complications 糖尿病是现代疾病中的第二杀手，其对人体的危害仅次于癌症。糖尿病本身并不可怕，可怕的是糖尿病的并发症。近年来，糖尿病发病率在全球呈增长趋势，在发展中国家显得尤为突出，我国2型糖尿病患者人数的增加同样触目惊心。作为影响患者生活质量的主要因素，糖尿病并发症长期困扰广大糖尿病患者。英国糖尿病前瞻性研究发现有半数新诊断的2型糖尿病患者中己存在一种或几种糖尿病慢性并发症。在一些国家，30%的慢性肾功能衰竭、40%~50%的失明、50%的心脑血管疾病、60%的截肢是由糖尿病造成的[1-4]。据估计，患有并发症的糖尿病人其经济负担是无并发症患者的数倍甚至几十倍。由此可见，糖尿病并发症的早期预测及诊断的实现将极大地改善糖尿病患者的生活质量，并减轻其庞大的经济开支。 本文基于糖尿