How to curve fit 4D data?

Hello :D

I am using Surface fitting toolbox to curve fit 3 dimesional data. Until now, I have 3 variables (X, Y and Z), and i can fit with no problems, and avaliate the goodnes of fit.

But now, I need to put there one more variable (T), and will be 4 dimensional. Any ideia how to fit that to obtain a mathematical expression, and avaliate the goodnes of fit?

Thank you

Answer 1

[Ahmet Cecen]

If your curves will be polynomials, try my FEX function MultiPolyRegress. It also provides a fairly rich set of goodness of fit measures.

http://www.mathworks.com/matlabcentral/fileexchange/34918-multivariate-polynomial-regression

Answer 2

[Sean de Wolski]

Statistics Toolbox provides fitnlm which will fit in N-D for you.

[Ahmet Cecen]

Or fitlm apparently for linear regression. When did this function get implemented? Seems my function is becoming obsolete (at least for people with the toolbox), except for the LOOCV trick, which is simple enough to add, wonder why they haven't.

[Sean de Wolski]

These are newer functions that provide similar functionality to nlinfit or mvregress

[Luis Meneses]

I have matlab R2011a, how can i do with this version?