Erratum to: A linear classifier based on entity recognition tools and a statistical approach to method extraction in the protein-protein interaction literature

Table 7 Performance of top 20 reported runs for the ACT in BC3

Team	Run	Acc.	Rank	F ₁	Rank	MCC	Rank	AUCiP/R	Rank	RP4
T73	RUN_2	.8915	1	.6132	2	.55306	1	.6796	2	4
T73	RUN_4	.8888	3	.6142	1	.55054	2	.6798	1	6
T73	RUN-1	.8755	16	.6083	3	.53524	3	.6591	3	432
T73	RUN_3	.8778	13	.6014	6	.52932	6	.6589	4	1872
T73	RUN_5	.8762	15	.6033	5	.53031	5	.6537	5	1875
T90	RUN_3	.8832	9	.5964	8	.52914	7	.6524	6	3024
T65	RUN_2	.8793	12	.5982	7	.52727	11	.6389	7	6468
T100	RUN_2	.8827	10	.5949	10	.52732	10	.6186	12	12000
T89	SRV_8	.8687	19	.6080	4	.53336	4	.4740	44	13376
T90	RUN_4	.8893	2	.5744	14	.52237	12	.4926	42	14112
T90	RUN_2	.8870	6	.5901	11	.5289	8	.5165	36	19008
T90	RUN-1	.8873	5	.5873	12	.52736	9	.5114	38	20520
T100	RUN-1	.8877	4	.5415	28	.50005	16	.6162	13	23296
T65	RUN_5	.8800	11	.5689	16	.50255	15	.6239	10	26400
T65	RUN-1	.8868	7	.5083	38	.48297	20	.6385	8	42560
T90	RUN_5	.8860	8	.5829	13	.52204	13	.5083	40	54080
T89	RUN_5	.8727	18	.5958	9	.52082	14	.4847	43	97524
T100	RUN_4	.8185	37	.5604	20	.4827	21	.6375	9	139860
T81	VTT3-SP	.8265	33	.5682	17	.49065	19	.5879	17	181203
T81	VTT1-SP	.8213	35	.5667	18	.49089	18	.5843	18	204120

The values obtained on the official BC3 gold standard using the F-Score, Accuracy, Matthew’s Correlation Coefficient, and Area Under the interpolated Precision and Recall Curve (computed with the official script, adding F-Score), as well as their respective ranks. RP4 denotes the rank product of these 4 measures. Boldfaced values represent best and second-best performance values for each measure. Our two best runs are shown at the bottom of the table according to the RP4 measure these runs are ranked 19 and 20 among all runs submitted. Overall, our team (81) ranks 6^th among all participating teams.

ISSN: 1471-2105