首页 | 本学科首页   官方微博 | 高级检索  
     检索      

多元回归分析方法应用实证分析与比较
引用本文:白秀琴,李瑞阁.多元回归分析方法应用实证分析与比较[J].河南科学,2010,28(9):1083-1088.
作者姓名:白秀琴  李瑞阁
作者单位:1. 平顶山工业职业技术学院,河南,平顶山,467001
2. 南阳理工学院,应用数学系,河南,南阳,473006
基金项目:河南省教育厅自然科学研究计划项目 
摘    要:利用多元统计分析的线性回归分析理论,首先对糖尿病临床采集数据进行相关分析,了解因素之间的相关关系,其次分别采用线性回归的前进法、后退法、逐步回归分析方法,探讨糖尿病病人胰岛素每日注射初始剂量与其影响因素之间的线性关系.根据回归方程调整相关系数、AIC准则等,对比选择最优应用回归模型,确定胰岛素用量及其主要影响因素血糖、尿蛋白的回归模型,对医生临床治疗提供参考.

关 键 词:多元回归分析  正态性检验  相关分析  前进法回归  后退法回归  逐步回归

Applied Comparison of Multiple Regression Analysis in the Practical
Bai Xiuqin,Li Ruige.Applied Comparison of Multiple Regression Analysis in the Practical[J].Henan Science,2010,28(9):1083-1088.
Authors:Bai Xiuqin  Li Ruige
Institution:1.Pingdingshan Industrial College of Technology,Pingdingshan 467001,Henan China;2.Department of Applied Mathematics,Nanyang Institute of Technology,Nanyang 473006,Henan China)
Abstract:Using Multiple linear regression analysis theory in the multivariate statistical analysis.Firstly,to proceed correlation analysis by the data in the clinical work and conclude the correlation relations between the factors,secondly,adobt forward,backward and stepwise regression analysis to explore the linear relationships between preliminary insulin dosage about diabetes mellitus patients and its effect factors.According to the adjusted correlation coefficients,AIC criterion etc.The optimal applied regression models between the insulin dosage and its major influence factors,blood sugar and urinary protein were obtained.The study gives mathematical model and some guide to clinic treatment.
Keywords:multiple regression analysis  test of normality  correlation analysis  forward regression  backward regression  stepwise regression
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号