Table 1 Performance measures of runs submitted to IAS

From: Uncovering protein interaction in abstracts and text using a novel linear model and word proximity networks

	SVM	VTT		SVD-UI
	Run 1	Run 2	Run 2'	Run 3	Run 3'	Mean^a	StDev^a	Median^a
Total predictions			750
Total positive			375
Total negative			375
True positives (TP)	330	295	323	300			N/A
False positives (FP)	186	118	133	143
True negatives (TN)	189	257	242	232
False negatives (FN)	45	80	52	75
Precision	0.64	0.71	0.71	0.68		0.66	0.08	0.68
Recall	0.88	0.79	0.86	0.8		0.76	0.19	0.85
Accuracy	0.69	0.74	0.75	0.71		0.67	0.06	0.67
F-score	0.74	0.75	0.78	0.73		0.69	0.10	0.72
FP rate	0.5	0.32	0.36	0.38			N/A
TP rate	0.88	0.79	0.86	0.8
Error rate	0.31	0.26	0.25	0.29
AUC	0.8	0.76	0.8	0.71	0.75	0.74	0.07	0.75

^aCalculated from 51 runs submitted by 19 teams. AUC, area under the curve; IAS, interaction article subtask; SVD, singular value decomposition; SVM, support vector machine; SVD-UI, SVD with uncertainty integration; VTT, variable trigonometric threshold. Bold entries for accuracy, F-Score, and AUC denote best value obtained for all our submitted runs.

Back to article page

ISSN: 1474-760X

Contact us

Submission enquiries: editorial@genomebiology.com
General enquiries: info@biomedcentral.com