生长曲线分析中的多水平模型及算法  

The Multilevel Model in Growth Curve Analysis and Programs

作　　者：王晓东[1] 田俊[2] 

机构地区：[1]泉州师范学院数学与计算机科学学院,福建泉州362000 [2]福建医科大学公共卫生学院,福建福州350004

出　　处：《数学的实践与认识》2011年第16期48-58,共11页Mathematics in Practice and Theory

基　　金：国家自然科学基金(81041066)

摘　　要：生长曲线分析是了解事物随时间的变化特点的分析热点.然而,传统的曲线拟合方法不适用通过重复测量得到的结构资料.采用轮廓设计矩阵运算虽然解决结构资料的生长曲线分析,但是相应的SAS程序冗长且难懂.将多水平模型应用于生长曲线的分析,通过实例分析介绍应用过程,同时给出简单明了的SAS程序.The growth curve analysis is a hotspot in the analysis to understand the characteristics of things change over time. However, the conventional curve fitting method does not apply to the structure data obtained by repeated measurement. Although the outline design matrix computation can be used to treat the structure data in the growth curve analysis, the corresponding SAS programs are very long and hard to understand. In this paper, a multi-level model will be applied to the analysis of the growth curve. The usage of the method is demonstrated by practical examples. The very simple and clear SAS programs for the presented method are also given.

关 键 词：生长曲线 多层结构数据 多水平模型 算法 

分 类 号：R195[医药卫生—卫生统计学]