Basis function custom problem in fitrgp()

8 次查看(过去 30 天)
I used hfcn = @(x)[v1,x]; to custom my own basis function when I fit Gaussian Process Regression using fitrgp(x,y,'BasisFunction',hfcn,'Beta',beta0,'KernelFunction','ardsquaredexponential','Verbose',1, ... 'Optimizer','fminunc', .... However, I was returned with an error "Error using cat. Dimensions of matrices being concatenated are not consistent." I am pretty sure that my v1 is of dimension n by 1 and my x used in fitrgp() is of dimension n by 1 as well. Could you please let me know anything wrong?

回答(1 个)

Hyunjung Lee
Hyunjung Lee 2019-7-21
What is the dimension of beta0? I would suggest that you check if beta0 is a 2 by 1 matrix since your basis is n by 2.

Community Treasure Hunt

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

Start Hunting!

Translated by