optimizing GPR model using genetic algorithm

As I tried to optimize the hyperparameters and sigma using GA, it is showing some error. Also, is it possible to limit the maximum

objective evaluations or put lb/ub on GA optimization? Please help.

clear;clc;close all

load('data001.mat')

x=data001(:,1); 
y=data001(:,2);

kernel = optimizableVariable('KernelFunction',{'exponential','squaredexponential','matern32','matern52',...
'rationalquadratic','ardexponential','ardsquaredexponential','ardmatern32','ardmatern52','ardrationalquadratic'},...
'Type','categorical');

sigma = optimizableVariable('Sigma',[1e-4,10],'Transform','log');

go = ga(@objFcn,2)  % calling objective function in optimization algorithm

kernel= go.XAtMinObjective.KernelFunction

model = fitrgp(x,y, 'KernelFunction', char(kernel));

yfit = predict(model, x);


function Loss = objFcn(Vars, x, y)
m = fitrgp(x, y, 'KernelFunction', char(Vars.KernelFunction), ...
                 'Sigma', Vars.Sigma, 'ConstantSigma', true,...
                 'KFold', 5);
Loss = kfoldLoss(m);
end

[Alan Weiss]

You asked this question elsewhere. See my answer here: https://www.mathworks.com/matlabcentral/answers/1754650-changing-optimization-technique-for-gaussian-process-regression-model?s_tid=srchtitle

Alan Weiss

MATLAB mathematical toolbox documentation