Confidence Regions Around the Ridge of Optimal Response on Fitted Second-Order Response Surfaces

Ridge analysis is a technique useful in connection with the fitting of second-order response surfaces. It enables a path of maximum predicted ŷ values (maximum on a sphere of radius R) to be specified, and followed out from the origin of the experimental region, as R increases. (Alternatively, a minimum ŷ path can be followed out.) This article discusses the addition of a measure of uncertainty to the selected path, thus providing an assessment of the reliability of the ridge analysis as a path progresses. The utility of this method is illustrated by applications to two datasets.

[1]  George E. P. Box,et al.  A CONFIDENCE REGION FOR THE SOLUTION OF A SET OF SIMULTANEOUS EQUATIONS WITH AN APPLICATION TO EXPERIMENTAL DESIGN , 1954 .

[2]  Arthur E. Hoerl,et al.  Application of ridge analysis to regression problems , 1962 .

[3]  Norman R. Draper,et al.  "RIDGE ANALYSIS" OF RESPONSE SURFACES , 1963 .

[4]  R. H. Myers,et al.  Response Surface Techniques for Dual Response Systems , 1973 .

[5]  W. H. Carter,et al.  Confidence regions for constrained optima in response-surface experiments. , 1983, Biometrics.

[6]  Roger Hoerl,et al.  Ridge Analysis 25 Years Later , 1985 .

[7]  Roger Hoerl.  The application of ridge techniques to mixture data: ridge analysis , 1987 .

[8]  U. Böckenholt,et al.  Analyzing optima in the exploration of multiple response surfaces. , 1989, Biometrics.

[9]  John J. Peterson.  A General Approach to Ridge Analysis With Confidence Intervals , 1993 .

[10]  Raja Aravamuthan,et al.  Application of Response Surface Methodology for the Maximization of Concora Crush Resistance of Paperboard , 1993 .

[11]  Enrique Del Castillo,et al.  THE COMPUTATION OF GLOBAL OPTIMA IN DUAL RESPONSE SYSTEMS , 1997 .

[12]  J. Semple,et al.  Optimality conditions and solution procedures for nondegenerate dual-response systems , 1997 .

[13]  Enrique del Castillo,et al.  Optimization of dual response systems: A comprehensive procedure for degenerate and nondegenerate problems , 1999, Eur. J. Oper. Res..

[14]  Steven G. Gilmour,et al.  Controlling processes in food technology by simplifying the canonical form of fitted response surfaces , 1999 .

[15]  Steven G. Gilmour,et al.  Some practical advice on polynomial regression analysis from blocked response surface designs , 2000 .

[16]  Norman R. Draper,et al.  Generalized Ridge Analysis Under Linear Restrictions, With Particular Applications to Mixture Experiments Problems , 2002, Technometrics.