Algorithm 1: Pseudo-code of NMI-TS algorithm

Input: dataset {P, Y} Results: new dataset A after selection

While P ≠ ∅ or S ≠ ∅   release the solution that reaches the tabu length;   j = 1;   While j <= the length of P     ${\mathrm{B}}_{\mathrm{j}}=\mathrm{I}\left(\mathrm{Y};{\mathrm{p}}_{\mathrm{j}}\right)-\mathsf{\xi }\frac{1}{\mathrm{k}}{\displaystyle {{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{k}}}\mathrm{I}\left({\mathrm{a}}_{\mathrm{i}};{\mathrm{p}}_{\mathrm{j}}\right);$     J = j + 1;   end   Sort the variables in P by Bj from largest to smallest and select the first N variables, put them in set M;   N = 1;   While n <= N     Evaluate λn for nth variable in M; (by using Algorithm 2)   n = n + 1;   end   Select the variable corresponding to the smallest λn and add it to A, Put the other variables in S; End