Table 5.

System performance for plain entity recognition on the MEDLINE test corpus. Data shown in italic font presents runs that were submitted after the official deadline. The median and average are computed solely using the official runs. A * symbol indicates statistically significant difference of a run with the runs ranked before and after it, per student test.

Team	TP	FP	FN	Precision	Recall	F-measure
Erasmus-run3.unofficial*	2220	1045	881	0.680	0.716	0.698
Erasmus-run1*	2139	1330	962	0.617	0.690	0.651
Erasmus-run2*	2103	1273	998	0.623	0.678	0.649
SIBM-run2*	1357	761	1745	0.641	0.438	0.520
SIBM-run1*	1476	1258	1626	0.540	0.476	0.506
BITEM-run1*	1376	1032	1741	0.571	0.442	0.498
LITL-run1*	998	556	2105	0.642	0.322	0.429
LITL-run2	989	561	2114	0.638	0.319	0.425
UPF-run2.unofficial*	969	5050	2138	0.161	0.312	0.212
UPF-run1*	736	5053	2369	0.127	0.237	0.166
UPF-run2	739	5050	2367	0.128	0.238	0.166
average				0.503	0.426	0.446
median				0.617	0.438	0.498