Table 1 Performance measures of the six different classifiers

Model	Dataset	SENS	SPC	ACC	PREC	NPV	F1	LRP	LRN
naiveBayes	orig	0.6991	0.4186	0.6218	0.7596	0.3462	0.7281	1.2025	0.7188
	marker	0.8000	0.3874	0.5064	0.3462	0.8269	0.4832	1.3059	0.5163
	pca	0.6731	0.3365	0.4487	0.3365	0.6731	0.4487	1.0145	0.9714
kNN	orig	0.6711	0.5000	0.6667	0.9808	0.0385	0.7969	1.3421	0.6579
	marker	0.6688	0.5000	0.6667	0.9904	0.0192	0.7984	1.3377	0.6623
	pca	0.6643	0.3077	0.6346	0.9135	0.0769	0.7692	0.9596	1.0909
logReg	orig	0.6923	0.3846	0.5897	0.6923	0.3846	0.6923	1.1250	0.8000
	marker	0.6615	0.3077	0.6026	0.8269	0.1538	0.7350	0.9556	1.1000
	pca	0.6623	0.0000	0.6538	0.9808	0.0000	0.7907	0.6623	Inf
randomForest	orig	0.6929	0.4483	0.6474	0.8462	0.2500	0.7619	1.2559	0.6850
	marker	0.6923	0.4103	0.6218	0.7788	0.3077	0.7330	1.1739	0.7500
	pca	0.6748	0.3636	0.6090	0.7981	0.2308	0.7313	1.0604	0.8943
rpart	orig	0.7083	0.4722	0.6538	0.8173	0.3269	0.7589	1.3421	0.6176
	marker	0.7404	0.4808	0.6538	0.7404	0.4808	0.7404	1.4259	0.5400
	pca	0.6935	0.4375	0.6410	0.8269	0.2692	0.7544	1.2330	0.7005
SVM	orig	0.6645	0.0000	0.6603	0.9904	0.0000	0.7954	0.6645	Inf
	marker	0.6623	0.0000	0.6538	0.9808	0.0000	0.7907	0.6623	Inf
	pca	0.6623	0.0000	0.6538	0.9808	0.0000	0.7907	0.6623	Inf

Performance measures of the six different classifiers on the different input data: orig-full data set; marker- candidate marker proteins only; pca–PCA transformed data. The meaning of the columns: SENS-sensitivity, SPC-specificity, ACC-accuracy, PREC-precision (positive predictive value), NPV-negative predictive value, F1-F-measure, LRP- likelihood ratio positive, LRN-likelihood ratio negative.

ISSN: 1471-2415