Skip to main content

Table 8 The leave-one-out and 5-fold cross validation classification accuracies of the SVM-classifier and the KNN-classifier based on four gene selection methods, GS1, GS2, Cho's, and F-test, on the PROSTATE dataset.

From: A stable gene selection in microarray data analysis

PROSTATE 5-Fold KNN SVMs
  30 60 100 Best Accuracy/# Genes 30 60 100 Best Accuracy/#Genes
GS2 0.917 ± 0.073 0.916 ± 0.056 0.913 ± 0.057 0.921/39 0.884 ± 0.080 0.908 ± 0.057 0.909 ± 0.060 0.911/91
GS1 0.918 ± 0.073 0.917 ± 0.056 0.907 ± 0.062 0.922/35 0.887 ± 0.082 0.901 ± 0.060 0.914 ± 0.058 0.914/99
Cho's 0.870 ± 0.144 0.918 ± 0.055 0.914 ± 0.058 0.918/10 0.841 ± 0.149 0.890 ± 0.069 0.904 ± 0.061 0.904/4
F-test 0.921 ± 0.053 0.915 ± 0.056 0.913 ± 0.057 0.935/61 0.893 ± 0.060 0.907 ± 0.062 0.914 ± 0.058 0.918/92
PROSTATE LOO KNN SVMs
  30 60 100 Best Accuracy/# Genes 30 60 100 Best Accuracy/#Genes
GS2 0.931 0.922 0.922 0.941/8 0.902 0.902 0.941 0.951/47
GS1 0.931 0.922 0.902 0.951/8 0.931 0.912 0.922 0.951/49
Cho's 0.931 0.912 0.912 0.941/8 0.941 0.912 0.912 0.941/20
F-test 0.931 0.922 0.922 0.941/10 0.892 0.931 0.931 0.941/4