multivariable regression, known structure, calculate coefficients

2 次查看（过去 30 天）
Ansgar Elles
Ansgar Elles 2020-8-8
评论： the cyclist 2020-8-11
Hi, I have the equations for calculating the corrosion rate (r_corr) of different materials (e(1) steel, e(2) zinc, e(3) copper, e(4) aluminium, see below). (According to ISO 9223).
My goal is now to create a fifth equation for another type of material. I have generated data where r_corr was measured along P_d, RH (relative humidity), T and S_d.
How can you calculate the coefficients (marked below in yellow) for the new material? I want to use the structure of the equations below. So to speak:
r_corr = a * P_d^b * exp(c * RH + f_St) + d * S_d^e*exp(f * RH + g * T)
Thank you very much!

请先登录，再回答此问题。

回答（1 个）

the cyclist
the cyclist 2020-8-8
If you have the Statistics and Machine Learning Toolbox, you should be able to use the fitnlm function.
  2 个评论
the cyclist
the cyclist 2020-8-11
I'm not too surprised. With 7 free parameters, you'll like need a lot of data to fit a stable statistical model.
If you can post your data, it might help people have some more specific ideas.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Linear Regression 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by