Table 3.

Overall results of the different configurations of BoB's false positives filtering component: instance-level results with fully contained matches and using the testing corpus

BoB's false positives filtering component	BoB individual classifiers	BoB multi-class	BoB best configuration
Overall micro-averaged (PHI level)
Precision	0.895	0.952	0.836
Recall	0.895	0.886	0.922
F1 measure	0.895	0.918	0.877
F2 measure	0.895	0.898	0.904

PHI, protected health information.