How to use Kernel Ridge Regression Function in Matlab - fitrkernel

I am trying to perform regression using kernel ridge regression in Matlab. Is there a dedicated function for Kernel Ridge Regression in Matlab?

I used the fitrkernel function, but the result is not correct. Am I doing something wrong?

Secondly, how can I specify the kernel to use.

N = 100;
data_x = linspace(0,20,N);
data_x = data_x';

data_y = sin(data_x);

[Mdl, FitInfo] = kernel_ridge_regression1(data_x, data_y);


x_test = linspace(0,15,75);
x_test = x_test';

actual_y = sin(x_test) ;


y_predict = predict(Mdl,x_test); % I used same data to test the results




figure(2)
plot(x_test,actual_y, 'o','lineWidth',2);
hold on
plot(x_test,y_predict, 'r','lineWidth',2);

legend("Original y", "Predicted y", 'interpreter','latex')
xlabel ('$x$','fontsize',18,'interpreter','latex')
ylabel('$y$','fontsize',18,'interpreter','latex')

%% FUNCTION

function [Mdl, FitInfo] = kernel_ridge_regression1(data_x_train, data_y_train)


    data_x_train_b = zscore(data_x_train); % Standardize the data

    [Mdl,FitInfo] = fitrkernel(data_x_train_b ,data_y_train);




end

Answer 1

[Anjaneyulu Bairi]

I Understand that you are trying to perform regression using kernel ridge regression. Try to explore the below link and download the files which has example and explanation of Kernel Ridge Regression and please note that it is a third-party support package and not provided by MathWorks.

• https://www.mathworks.com/matlabcentral/fileexchange/49989-kernel-ridge-regression-in-matlab

I hope it helps to resolve your query.