Polynomial MLR to fit multiple compositional values to single value for large data set?

I'm trying to fit 8 compositional values to a single performance value measured with an analyzer for a data set of 100 samples. MLR isn't quite getting me there and I was wondering if adding some square terms might get me closer. Are there any toolkits, apps or code that I could try?

[Larry McDermott]

OK...that was kind of dumb, I redid my x-matrix to add square, cubic and square root values of my originals, but if there are some apps that can include terms like this including selection rules I'd like to give it a shot!

Answer 1

[Siriniharika Katukam]

You can try using custom models in curve fitting toolbox under linear and non-linear regression to define custom equations. The link below might be helpful:

https://www.mathworks.com/help/curvefit/custom-nonlinear-models.html