| 1: Inputs:
X,Y and number genes (r) to be selected. |
| 2: Output: Sequence of selected genes T. |
| 3: for all
j ∈ H do
|
| 4: for
c = 0,1 do
|
| 5: Compute the relative dominant class for each gene, i.e., R(j,c) in Eq. (1). |
| 6: end for
|
| 7: for
i→N
do
|
| 8: Compute the gene mask for each gene, i.e., mji as defined in Eq. (7). |
| 9: Compute the RPOSjscores for each gene as defined in Eq. (9). |
| 10: Assign RDCJ to each gene as defined in Eq. (8). |
| 11: end for
|
| 12: let M ∈ P×N be the gene mask matrix M = [mji], where its ithvalue for jthgene is either 0 or 1. |
| 13: Compute the total or aggregate mask of genes and denote it by M..(H). |
| 14: Use the Greedy search approach to select the minimum subset of genes from M, M..(H) and RPOSj and denote it by H*. |
| 15: Perform H = H − H*, this will exclude the genes selected in minimum subset from the whole set of genes. |
| 16: Arrange the genes in RDCjin the increasing order of RPOSJ for each class. |
| 17: end for
|
| 18: Obtaining final listed or ranked genes.
|
| 19: if
r ≤ |H*| then
|
| 20: Then T includes the genes which are first r genes in H*. |
| 21: while |T| < r
do
|
| 22: Increase T by one gene in a round-robin fashion method. |
| 23: end while
|
| 24: end if
|
| 25: return
T
|
