Algorithm 1.

sigFeature

Inputs:       Training examples $X_0={\left[x_1,x_2,\dots ..x_k,\dots x_l\right]}^T$       Class labels $y={\left[y_1,y_2,\dots ..y_k,\dots y_l\right]}^T$ Initialize: Subset of existing features s = [1, 2, …n] Feature ranked list r = [] Repeat until s = []            Confine training examples to good feature inventories           X = X0(:, s)           Train the classifier (training data set)           α = SVM − train (X, y)           Enumerate the weight vector of dimension length(s)           $W=\sum _k{\alpha }_k{y}_k{x}_k$           Enumerate the ranking criteria (for all i)           ci = wi*P[ti(v) < −|u|orti(v) > |u|]           Find the feature with the negligible ranking criterion           f = sort(c)           Reform feature ranked list:           r = [s(f), r]           Remove the feature with the smallest ranking criterion           s = s(1:f − 1, f + 1:length(s)) Output: Feature ranked list r.