Rule discovery and distance separation to detect reliable miRNA biomarkers for the diagnosis of lung squamous cell carcinoma

Table 2 Comparisons of three classifiers on four data sets.

Data sets	Algorithms	Specificity	Sensitivity	F-Measure	ROC Area
15 miRNAs (Raponi et al. 2009)	KNN	0.9833	0.8182	0.975	0.934
	NB	0.9833	0.8182	0.975	0.934
	C4.5	0.9516	0.7778	0.959	0.827
5 miRNAs (Shen et al. 2010)	KNN	0.9839	1.0000	0.992	0.944
	NB	0.9839	1.0000	0.992	0.989
	C4.5	0.9672	0.8000	0.967	0.84
19 miRNAs (top ranked)	KNN	0.9839	1.0000	0.992	0.944
	NB	0.9836	0.9000	0.984	0.946
	C4.5	0.9524	0.8750	0.968	0.798
309 miRNAs (lower ranked)	KNN	0.9833	0.8182	0.975	0.926
	NB	0.8413	0.6250	0.935	0.779
	C4.5	0.8413	0.3846	0.891	0.666

ISSN: 1471-2164