Table 2.

GeneId-Ranking Methods on Development Data

	Mean Average Precision (MAP)
mouse development data
likely-prot + softTFIDF	0.450
possible-prot + softTFIDF	0.626
graph-based ranking	0.513
+ extra links	0.730
+ extra links & learning	0.807

Mean average precision of several geneId-ranking methods on the 50 abstracts from the mouse development dataset. The first two lines are NER systems, coupled with using softTFIDF to matching extracted names with gene synonyms. The graph-based ranking method (see text) ranks geneIds by proximity to a query abstract in a graph that includes the results of both NER systems.