Selecting significant marker genes from microarray data by filter approach for cancer diagnosis

The purpose of classification is to predict the presence of cancer. The classification accuracy is measured for each selected subset of genes. Results show that the combination between S/R selection method and the KNN classifier present the highest accuracies for different cancers; 100% for only 13 relevant genes in Leukemia cancer, 100% for 21 genes in Lung cancer, 100% for 4 genes in Lymphoma cancer, 76.7% for 6 genes in CNS cancer, 100% for 30 genes in ovarian cancer.