kernel_functions
Rolul acestor kernel-uri este de a ne oferi o modalitate de a estima parametri necesari in functie de gradul maxim (din punct de vedere polinomial) pe care il atribuim functiei ϕ.
Kernels
Linear Kernel
function retval = linear_kernel(x, y, other)
% ignore 'other' parameter
retval = dot(x, y);
endfunction
Polynomial Kernel
function retval = polynomial_kernel (x, y, d)
% kernel polynomial
retval = (1 + dot(x, y))^d;
endfunction
Gaussian/Radial - Basis Kernel
function retval = gaussian_kernel (x, y, sigma)
squared_distance = sum((x - y).^2);
retval = exp(-squared_distance / (2 * sigma^2));
endfunction
Last modified: 30 April 2024