BMC Bioinformatics

Table 5 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 LUNG dataset.

From: A stable gene selection in microarray data analysis

LUNG 5-Fold	KNN				SVMs
	30	60	100	Best Accuracy/#Genes	30	60	100	Best Accuracy/#Genes
GS2	0.884 ± 0.053	0.916 ± 0.041	0.928 ± 0.037	0.928/100	0.858 ± 0.061	0.913 ± 0.035	0.931 ± 0.033	0.931/99
GS1	0.890 ± 0.046	0.919 ± 0.041	0.937 ± 0.034	0.937/99	0.871 ± 0.051	0.922 ± 0.038	0.938 ± 0.031	0.938/98
Cho's	0.843 ± 0.053	0.897 ± 0.044	0.924 ± 0.038	0.924/100	0.803 ± 0.065	0.894 ± 0.044	0.924 ± 0.035	0.924/100
F-test	0.873 ± 0.049	0.882 ± 0.044	0.918 ± 0.044	0.918/100	0.852 ± 0.055	0.901 ± 0.042	0.930 ± 0.036	0.930/100
LUNG LOO	KNN				SVMs
	30	60	100	Best Accuracy/# Genes	30	60	100	Best Accuracy/# Genes
GS2	0.892	0.906	0.921	0.931/44	0.867	0.892	0.931	0.931/73
GS1	0.887	0.941	0.941	0.951/49	0.862	0.941	0.941	0.951/51
Cho's	0.837	0.897	0.921	0.926/86	0.773	0.892	0.931	0.941/88
F-test	0.872	0.877	0.901	0.921/89	0.857	0.901	0.926	0.936/94

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com