function y = Sparse_L_ij(num_columns,npp,N,l5_rec,l23_rec,l5_l23,l23_l5,l23_l23_cross,l5_l5_cross,noise)
%instantiation of weight matrix
L_ij1 = zeros(N,N);
rho = .3;%.0675;
for i = 1:num_columns
    j = (i-1)*npp*2;
    for k = 0:2*npp:(num_columns - 1)*2*npp
        if k ~= j
            L_ij1(j+npp+1:j+(2*npp),k+1:k+npp) = l23_l5*spones(sprandn(npp,npp,rho));
            L_ij1(j+npp+1:j+2*npp,k+npp+1:k+2*npp) = l23_l23_cross*spones(sprandn(npp,npp,rho));
            L_ij1(j+1:j+npp,k+1:k+npp) = l5_l5_cross*spones(sprandn(npp,npp,rho));
        end
    end
    L_ij1(j+1:j+npp,j+1:j+npp) =  l5_rec*spones(sprandn(npp,npp,rho));
    L_ij1(j+npp+1:j+2*npp,j+npp+1:j+2*npp) =  l23_rec*spones(sprandn(npp,npp,rho));
    L_ij1(j+1:j+npp,j+npp+1:j+2*npp) =  l5_l23*spones(sprandn(npp,npp,rho));
end
L_ij1 = L_ij1 + noise*randn(N).*(L_ij1>0);
y = abs(transpose(L_ij1(1:N,1:N)));
end