An improved machine learning protocol for the identification of correct Sequest search results

BMC Bioinformatics

Table 2 Validation metrics for a collection of machine learning algorithm runs over testsets containing feature from the groups denoted in table 1

Feature groups	Algorithm	Accuracy	Sensitivity	Specificity	AUC ROC	Net pred.
All	ABWillow	0.97505	0.56504	0.99385	0.96379	0.77945
	ABC4.5	0.97361	0.58815	0.99269	0.94821	0.79042
	RFC4.5	0.97276	0.57212	0.99259	0.87901	0.78235
	ADtree	0.97688	0.7248	0.98988	0.96923	0.86118
SEQUEST	ABWillow	0.96951	0.48762	0.99336	0.90723	0.74050
	ABC4.5	0.97258	0.57018	0.99250	0.907084	0.78139
Published	RFC4.5	0.97228	0.58961	0.99122	0.912744	0.79042
	ADtree	0.96925	0.48762	0.99310	0.90604	0.74032
-	PeptideProphet	0.9688	0.54	0.99	-	0.765

ISSN: 1471-2105