Forensic Speaker Comparison Using Evidence Interval in Full Bayesian Significance Test

Table 2

Percentage of true positives, true negatives, false positives, and false negatives obtained in the test and validation steps.

		True positive (%)	True negative (%)	False positive (%)	False negative (%)	In dubio pro reo (%)
Train	Punctual	91.9	91.9	8.1	8.1
	Gosset	91.1	90.8	7.0	7.4	1.8
	Morrison	91.2	90.9	7.1	7.5	1.7
	FBST	90.8	90.0	6.79	7.12	2.4
Validation	Punctual	85.3	90.6	9.4	14.7
	Gosset	85.0	90.0	8.8	14.0	1.1
	Morrison	85.0	90.0	8.8	13.9	1.2
	FBST	85.0	89.9	8.7	13.8	1.3