多元非线性回归参数估计的数论方法及应用  

Number-theoretic Method for Parameters Estimation in Multivariate Nonlinear Regression with Application

作　　者：孙晓武[1,2] 方积乾[1,2] 杨杏芬[1,2] 

机构地区：[1]中山医科大学卫生统计学教研室 [2]中山医科大学卫生毒理学教研室

出　　处：《中国卫生统计》1999年第1期6-7,共2页Chinese Journal of Health Statistics

基　　金：中华医学基金

摘　　要：目的构造多元非线性回归模型，分析职业流行病学研究中职业暴露的多个效应。方法应用数论方法解决模型的参数估计问题。结果在二硫化碳对工人的健康评价中，建立了二元非线性回归方程，数论方法所得参数估计的标准误小于ＳＡＳＰｒｏｃＮＬＩＮ一元回归结果。结论数论方法为多元非线性模型提供了简便易行的参数估计手段。Objective Developing multivariate nonlinear regression models to characterize the effect of occupational exposure in occupational epidemiology.Methods Number-theoretic(NT)method was used to estimate parametes in model.Results In the evaluation of health effects of CS 2,two-outcome nonlinear model was develped,and the standard errors of estimated parameters from NT-method were smaller than those from SAS Proc NLIN.Conclusion NT method provides a simple and easy way to the estimation for multivariate nonlinear models.

关 键 词：多元非线性回归 数论方法 职业流行病学 CS2 

分 类 号：R181.31[医药卫生—流行病学] R195[医药卫生—公共卫生与预防医学]