陕西省汉中市农村单纯性收缩期高血压影响因素的分类树研究
目的探索陕西省汉中市农村地区单纯性收缩期高血压(is01atedsystolichypertension,IsH)的影响因素,为IsH的预防和控制提供依据和建议。方法以陕西省汉中市农村居民健康调查资料为数据源,采用分类树(CART)分析方法拟合模型,选择适合二分类变量的cART算法,并以错分率、正确分类率和R0C曲线等指标评价模型。结果该模型的训练集、验证集和测试集的错分率分别为16.6%、16.8%、14.3%,RootAsE分别为O.357、O.351、0.334,R0C曲线下面积大于O.5,说明模型拟合得较好。该模型首先按年龄将全人群分裂,影响因素按重要性依次为年龄、体质指数、受教育水平...
Saved in:
Published in | 西安交通大学学报(医学版) Vol. 34; no. 3; pp. 403 - 405 |
---|---|
Main Author | |
Format | Journal Article |
Language | Chinese |
Published |
西安交通大学医学院公共卫生系,陕西西安,710061%汉中市人民医院心内科,陕西汉中,723000
2013
|
Subjects | |
Online Access | Get full text |
ISSN | 1671-8259 |
Cover
Summary: | 目的探索陕西省汉中市农村地区单纯性收缩期高血压(is01atedsystolichypertension,IsH)的影响因素,为IsH的预防和控制提供依据和建议。方法以陕西省汉中市农村居民健康调查资料为数据源,采用分类树(CART)分析方法拟合模型,选择适合二分类变量的cART算法,并以错分率、正确分类率和R0C曲线等指标评价模型。结果该模型的训练集、验证集和测试集的错分率分别为16.6%、16.8%、14.3%,RootAsE分别为O.357、O.351、0.334,R0C曲线下面积大于O.5,说明模型拟合得较好。该模型首先按年龄将全人群分裂,影响因素按重要性依次为年龄、体质指数、受教育水平、吸烟、家庭月平均收入和多盐。结论ISH受多因素影响,应针对其影响因素做好一级预防,对不同的高危人群给予针对性的干预策略,降低ISH的发病率,提高农村居民的生活质量。 |
---|---|
Bibliography: | FANG Shi-hua, YAN Hong , DANG Shao-nong , LI Qiang , ZHAO Ya-ling, LIU Xiao-ning, YANG Rui-hai , REN Yong (1. School of Public Health, Medical School of Xi'an Jiaotong University, Xi'an 710061~ 2. the People's Hospital of Hanzhong city, Hanzhong 723000, China) isolated systolic hypertension (ISH); classification and regression tree (CART); influencingfactor; rural areas in Hanzhong city; health investigation Objective To explore factors affecting isolated systolic hypertension (ISH) in rural areas of Hanzhong city, Shaanxi Province, so as to provide evidence for ISH prevention and control. Methods We used the data collected from a survey on people's health in rural areas in Hanzhong City, Shaanxi Province, as data resource in this research. Classification and regression tree (CART) fitting binary variable was selected and classification tree was used to fit the model. Misclassification rate, correct classification rate and ROC curve were used to evaluate the model. Results Misclassification rate of the model tr |
ISSN: | 1671-8259 |