Module Library

QUADREG Call

performs quadratic response surface regression

RUN QUADREG( xopt, yopt, type, parms, , );

The inputs to the GPROBCNT subroutine are as follows:

xopt

is a returned value containing critical factor values.

yopt

is a returned value containing the critical response value.

type

is a returned character string containing the solution type (maximum or minimum).

parms

is a returned value containing the parameter estimates for the quadratic model.

is an factor matrix, where is the number of factor variables and is the number of data points.

is an response vector.

The QUADREG module fits a regression model with a complete quadratic set of regressions across several factors. The estimated model parameters are divided into a vector of linear coefficients and a matrix of quadratic coefficients to obtain critical factor values that optimize the response. It further determines the type of the optima (maximum, minimum, or saddle point) by computing the eigenvalues of the estimated parameters.