Gaussian Process Regression:I can plot the graphs using "fitrgp", but can I get the mean μ(x) and standard deviation sigma�(x)?

I can plot the graphs using "fitrgp", but can I get the mean μ(x) and standard deviation sigma(x)?

Answer 1

[mizuki]

Yes. The first output argument of "resubPredict" is the mean and the second is the sigma. If you want to draw them, use errorbar.

% create samples
n = 10;
x = linspace(0.5,2.5,n)';
y = sin(10*pi.*x) ./ (2.*x)+(x-1).^4 + 1.5*rand(n,1);

% Gaussian Process Regression
gprMdl = fitrgp(x,y);
[ypred,ysd,yint] = resubPredict(gprMdl,'Alpha',0.01);

% draw samples and mean&variance
figure;
plot(x, y, 'o');
hold on;
errorbar(x, ypred,ysd)