多元线性回归程序及其在医学统计中的应用

Program of Multiple Liner Regression and Its Application in Medical Statistics Teaching

在医学统计课堂教学中将教材中比较繁复的多元线性回归内容借助于3M投影仪显示在学生面前将会有助于提高教学学习质量。出于上述目的,我们用CA—Realizer语言编制了该软件。该软件包括以下两大部分:1.课堂教学内容:内容提要;基本要求;多元线性回归问题的提出;多元线性回归模型;参数估计及预报方程;最小二乘法估计量的性质;预报方程意义的检验;预报方程以及预报值的可信限等。2.计算一个具体实例的多元线性回归方程:输入样本资料阵;计算出正规方程增广阵,及逆阵和回归系数向量;输出回归方程;方差分析表及给一具体独立变量集合,计算并输出预报值及可信限等。本程序可对维数1<=P<=30,任意多个样本进行回归计算(标号为BBB)。

During medical statistics's class' teaching, the complex multiple liner regression will be shown to the students by 3M projector and this way will contribute to increase the teaching - quality. According to the aim mentioned, we edit the software with CA-Realizer language. This software includes the following two sections; the first is the contents of the class - teaching. They include the contents'abstract, base asking, offering the question of multiple liner regression, multiple liner regression's model, parameter estimate, prediction equation, properties of the least aquare estimators, testing the meaning of prediction equation, prediction equation and the confidence limits of prediction equation etc. The second makes the sample of multiple liner regression equation. They include the data matrix's inputting, computing normal equation's sets, invert matrix, regression coefficient vector, out-putting regression equation, table of analysis of variance, giving independent variable's assemble and computing and outputting the prediction value and it's confidence limits etc. If dimention is between 1 and 30, this procedure may adapt to the random samples'regression compution.