BMC Genomics

Table 1 Evaluation measures for test data

From: How to evaluate performance of prediction methods? Measures and their interpretation in variation effect analysis

	-50 %	-25 %	Equal	+25 %	+50 %
tp	750	750	750	750	750
fn	250	250	250	250	250
tn	375	563	750	938	1125
fp	125	187	250	312	375
sensitivity	0.75	0.75	0.75	0.75	0.75
specificity	0.75	0.75	0.75	0.75	0.75
PPV	0.86	0.80	0.75	0.71	0.67
NPV	0.60	0.69	0.75	0.79	0.82
accuracy	0.75	0.75	0.75	0.75	0.75
MCC	0.48	0.50	0.50	0.50	0.49

Example of a situation when both positive and negative cases have the same rate of correct predictions (75 %) and when the amount of negative cases is either equal or 25 or 50 % lower or higher than that for positive cases.

Back to article page

ISSN: 1471-2164

Contact us

Submission enquiries: bmcgenomics@biomedcentral.com
General enquiries: ORSupport@springernature.com