Skip to main content

Table 3 Performance of gene feature selection methods with KNN classifier (high) and SVM classifier (low) in multiclass datasets

From: Feature selection of gene expression data for Cancer classification using double RBF-kernels

Dataset: Lymphoma
  DKBCGS GINI Χ2-Statistic Info.Gain KW RF MRMR KBCGS
ACC 1.0000 1.0000 1.0000 1.0000 0.8617 0.9756 1.0000 1.0000
Gene content 25 70 26 49 22 16 29 35
TPR 1.0000 1.0000 1.0000 1.0000 0.8617 0.9756 1.0000 1.0000
TIME(s) 0.3412 0.8944 2.3579 1.2561 7.4577 3.3144 1.5922 0.7541
Dataset: Lung cancer
  DKBCGS GINI Χ2-Statistic Info.Gain KW RF MRMR KBCGS
ACC 0.9554 0.9443 0.9499 0.9641 0.9273 0.9472 0.9291 0.9514
Gene content 32 82 97 65 39 88 50 40
TPR 0.9243 0.9033 0.9185 0.9012 0.9210 0.9123 0.9042 0.9155
TIME(s) 0.1215 0.2250 0.1954 0.1857 1.6478 0.5111 0.2931 0.3171
Dataset: Lymphoma
  DKBCGS GINI Χ2-Statistic Info.Gain KW RF MRMR KBCGS
ACC 1.0000 0.994 1.0000 1.0000 0.9283 0.9963 1.0000 1.0000
Gene content 35 34 34 16 28 17 27 40
TPR 1.0000 0.994 1.0000 1.0000 0.9283 0.9963 1.0000 1.0000
TIME(s) 0.3412 0.8944 2.3579 1.2561 7.4577 3.3144 1.5922 0.7541
Dataset: Lung cancer
  DKBCGS GINI Χ2-Statistic Info.Gain KW RF MRMR KBCGS
ACC 0.9151 0.9041 0.9115 0.9229 0.9102 0.9087 0.9199 0.9100
Gene content 64 87 75 89 71 60 77 74
TPR 0.9172 0.9005 0.9124 0.9285 0.9089 0.9114 0.9207 0.9122
TIME(s) 0.5736 1.8912 3.4551 2.4972 6.9322 4.1978 2.1207 1.0044