基于数据挖掘技术的妊娠期糖尿病危险因素相关性研究  

The study of correlation between risk factors of gestational diabetes mellitus based on data mining technology

作　　者：马瑾[1] 徐东雨[1] 孙颖[2] 

机构地区：[1]中国医科大学基础医学院计算机教研室,辽宁沈阳110000 [2]中国医科大学门诊部,辽宁沈阳110000

出　　处：《中国医学工程》2014年第9期5-6,共2页China Medical Engineering

摘　　要：目的进行妊娠期糖尿病与微量元素、体质量指数等因素相关性的研究,探讨使用决策树模型预测孕妇患妊娠期糖尿病的风险,为临床预防提供参考依据。方法收集某社区医院2010年的孕妇体检资料279例,采用C5.0算法构建决策树模型,并使用分析节点对其预测结果进行评估。结果使用建立的决策树模型测试样本,有90.15%的测试样本的预测值与实际值相符,准确率较高。根据建立的决策树模型生成规则集,规则表明妊娠期糖尿病与体质量指数、钙、铁、年龄等相关因素关系密切。结论采用C5.0算法构建决策树模型预测妊娠期糖尿病是可行的,本方法可对大样本数据进行数据挖掘,其预测结果具有一定的应用价值。【Objective】For gestational diabetes mellitus and trace elements, body mass index factor relationship, to explore the use of decision tree model in pregnant women with gestational diabetes risk applications, to provide reference for clinical prevention. 【Methods】Collected in a community hospital in 2010 physical examination data of 279 cases of pregnant women, using C5.0 algorithm to construct decision tree model, and use the analysis node evaluating the results of prediction.【Results】Using the decision tree model test sample, 90.15% of the test samples of the prediction values and actual value is basically the same, the accuracy rate is high. Based on the decision tree model to generate sets of rules, the rules that body mass index, calcium, iron, age and other related factors closely related to the gestational diabetes mellitus.【Conclusion】Using C5.0 algorithm to construct decision tree model of gestational diabetes is feasible, the method can mine large sample data, the prediction results have a certain degree application value.

关 键 词：妊娠期糖尿病 数据挖掘 决策树模型 C5.0算法 

分 类 号：R587.1[医药卫生—内分泌]