应用SAS解非线性回归问题

The Application of SAS/STAT on Nonlinear Regression

.应用SAS/STAT估计非线性回归模型中的参数.首先,通过变量代换,把可以线性化的非线性回归模型化为线性回归模型,并用普通最小二乘法、主成分分析法和偏最小二乘法求模型中的参数和回归模型.其次,通过改良的高斯—牛顿迭代法来估计Logistic模型和Compertz模型中的参数.

We use SAS/STAT program to estimate nonlinear regression model parameters. The first, we transform nonlinear regression model into linear model with transformed variables and estimate model parameters by used the methods of ordinary least squares, principal component analysis and partial least squares regression, from which the model parameters and the regression line itself can be estimated. The second, we use modified Gauss-Newton iterative procedure to estimate Logistic and Compertz model parameters.