Flow-chart of the entire process. First the data set is divided into training and test sets. Then in the first stage, FSMLP is used to select 20 important genes. In the next stage, again FSMLP is used to select 10 useful genes from the 20 genes selected in the first stage. Now, NERFCM is applied to cluster the 20 genes selected in the first stage considering only the training data. Six clusters are found using NERFCM. The results of NERFCM and the second stage of gene selection are combined to choose just seven genes with sufficient cancer-specific signatures to distinguish between the 4 types of SRBCTs. Now different classifiers such as MLP, SVMs, RBF nets are trained using the training set with the selected 7 genes. The trained system (MLP/RBF net/SVM) is then used for classification of blind test data.