Fractional polynomial response surfaces
by
Steven Gilmour
Queen Mary, University of London
Coauthors: Luzia Trinca (UNESP-Botucatu)

Second order polynomial models have been used extensively to approximate the relationship between a response variable and several continuous factors. However, sometimes polynomial models do not adequately describe the important features of the response surface. The use of fractional polynomial models is described in this talk. It is shown how the models can be fitted, an appropriate model selected and inference conducted. Polynomial and fractional polynomial models are fitted to a published data set, which illustrate that sometimes the fractional polynomial can give as good a fit and much better interpolations than the polynomial model. Consideration will also be given to design and some results on Bayesian optimal design will be reported.

Date received: September 19, 2003