Figure 10From: Entropy-based gene ranking without selection bias for the predictive classification of microarray dataScheme of the E-RFE algorithm. In the case of linear SVM, ; in the non linear case, , where Z ij = y i y j  K(x i , x j ). We chose R t = 100, n int = , and M t = 0.2. At the end of the algorithm F, will contain the complete ordered list of the variables, and R will be empty.Back to article page