function [G,centers,sigmas] = generate_rbfs(X, N)

rbf_number = 0; % number of generated rbfs

for i = 1:N

rbf_number = rbf_number + (2^i + 1)^2;

end

G = zeros(length(X), rbf_number);

centers = zeros(size(X,2), rbf_number);

sigmas = zeros(1, rbf_number);

k = 1;

max_x = max(max(X));

min_x = min(min(X));

for i = 1:N

iter_rbf_count = 2^i+1; % number of RBFs in current iteration

sigma = sqrt(max_x - min_x) / 2^(i-1);

% generating centers for rbfs

for j = 1:iter_rbf_count

for l = 1:iter_rbf_count

for m = 1:iter_rbf_count

for m2 = 1:iter_rbf_count

centers(1, k) = min_x + (max_x - min_x) / (iter_rbf_count-1) * (j-1); %Burayi generalize et.

centers(2, k) = min_x + (max_x - min_x) / (iter_rbf_count-1) * (l-1);

centers(3, k) = min_x + (max_x - min_x) / (iter_rbf_count-1) * (m-1);

centers(4, k) = min_x + (max_x - min_x) / (iter_rbf_count-1) * (m2-1);

G(:,k) = RBFIO(X, sigma, centers(:,k)') ;

sigmas(k) = sigma;

k = k + 1;

end

end

end

end

end