In the second phase of Response Surface Methodology, using the central composite design on input/ output data, I want to fit a second-order polynomial on the subarea to find the optimum combination. Is there any mathematical formulation or toolbox in Matlab that can give me the estimated coefficients? I should code this in Matlab and do not want to use Minitab or Design
Expert software and I have four input data which makes it to estimate 15 coefficients.
Read the following pages from MATLAB documentation:
https://www.mathworks.com/help/stats/design-of-experiments-1.html
https://www.mathworks.com/help/stats/improve-an-engine-cooling-fan-using-design-for-six-sigma-techniques.html
Determine B =(inv (X'X))*X'*Y directly in the matlab/scilab console by first feeding X (design matrix ) corresponding to your design and then the observation vector Y into the console.
Debopam Ghosh I thought this formulation can only be used for first-order polynomials but seems like as long as the metamodel is linear in coefficients, it can be applied. Thanks.
- Badges
- Science topic