Skip to main content
. Author manuscript; available in PMC: 2009 Mar 9.
Published in final edited form as: Pac Symp Biocomput. 2006:52–63.

Table 3.

Best mappings of the 413 GeneRIF texts against their corresponding abstract titles and sentences under different Dice coefficient thresholds T. T < 05 is not considered as an acceptable match.

matching T = 05 T = 0.6 T = 0.7 T = 0.8 T = 0.9 T = 1.0
the title 25.4% 2.32% 19.9% 16.9% 12.3% 9.69%
the last sentence 26.1% 24.5% 20.6% 16.2% 9.93% 2.42%
the penultimate sentence 8.96% 7.51% 5.08% 4.36% 3.63% 0.97%

other sentences 17.7% 16.5% 13.0% 9.24% 5.65% 12.2%

total matching 78.7% 71.7% 58.6% 46.7% 31.5% 14.3%
no matching 21.3% 28.3% 41.4% 53.3% 68.5% 85.7%