突变模型的泛涵回归分割与疾病干预的临界控制及其策略评价  

A Functionalized Regressive Division on Saltation Model and Critical Control and Approach Evaluation in Disease Intervention

作　　者：陈立功[1] 陈咏梅[1] 余松林[1] 施侣元[1] 

机构地区：[1]武汉同济医科大学公共卫生学院,武汉430030

出　　处：《中国公共卫生》2000年第12期1061-1064,共4页Chinese Journal of Public Health

基　　金：UNDP/WorldBank/WHO热带病研究与培训特别规划署的资助!项目编号 :930 2 57

摘　　要：根据泛函分析、分段回归和广义Chow检验等 ,提出了在几种常见非线性突变模型中确定最优分割点的泛函回归分割 (FRD)法 ,又称临界回归分析 (CRA) ,其核心是计算各临界回归模型的最小合并残差均方根。应用广义线性模型(GLM)和SAS的GLM程序对血吸虫病控制的经济学研究验证了CRA的可行性。结果提示FRD可以确定曲线有显著意义的拐点 ,是用数学方法研究疾病的临界控制、评价并修正干预措施的唯一依据 。This paper has advanced a method of Functionalized Regressive Division(FRD),or Critical Regression Analysis (CRA) to .determine the optimum dividing points or inflection points for the non-linear saltation models, according to the Functionalized Analysis, Parted Regression and Generalized Chow-test.The CRA's core is to compute a minimum root of combined mean of squared residuals of all critical regression models. A functionalized statistical analysis (FSA) has been done with the theory of General Linear Models (GLM) and the GLM procedure of SAS.This paper demonstrated a feasibility and practical application of the CRA with an economical study on schistosomiasis control. The CRA can help us to determine the infiection with significance. It is the only basis for studying on critical control of disease, evaluating and modifying the intervention approach with mathematical method. The CRA or FRD can be widely applied in other fields.

关 键 词：非线性突变模型 泛涵回归分割 疾病干预 

分 类 号：R1[医药卫生—公共卫生与预防医学]