Response Surface Regression in GRM

Select Response surface regression on the General Regression Models Startup Panel - Quick tab to specify response surface designs. In the subsequent Quick Specs dialog box, specify one or more continuous predictor variables, an optional blocking variable, and one or more dependent variables. Note that the program will not perform any rescaling or centering of polynomials; the polynomial transformations will be applied directly to the respective raw data values (see also Experimental Design for a detailed discussion of response surface designs).

Response surface designs are commonly used in industrial experimentation, to explore nonlinear relationships between continuous predictor variables and the dependent variables. Such relationships usually occur when the predictor variable values are set so that the dependent variable of interest is at its optimum (e.g., at the peak hardness, or peak quality). Response surface designs make it possible for you to fit quadratic surfaces to the observed dependent variable values, and to determine the expected (predicted) largest (or smallest) values for the dependent variables of interest.

For more information, refer to the Introductory Overview and GRM - Index.