Table 2

Comparison of different annotation methods in terms of macro F1
Stage range Number of terms SVMSpatial + Sparse SVMSparse SVMSpatial SVMGlobal
4-6 10 .5224 ± .0407 .5094 ± .0393 .4926 ± .0414 .4767 ± .0386
20 .4454 ± .0461 .4200 ± .0462 .4141 ± .0459 .3794 ± .0412
30 .3459 ± .0593 .3230 ± .0516 .3153 ± .0565 .2942 ± .0479
7-8 10 .5372 ± .0343 .5282 ± .0312 .5131 ± .0329 .5055 ± .0329
20 .3653 ± .0517 .3603 ± .0538 .3331 ± .0740 .3364 ± .0676
9-10 10 .5561 ± .0282 .5499 ± .0276 .5353 ± .0289 .5267 ± .0260
20 .3836 ± .0464 .3764 ± .0442 .3527 ± .0370 .3429 ± .0342
11-12 10 .6339 ± .0280 .6261 ± .0269 .6109 ± .0271 .6060 ± .0257
20 .5226 ± .0379 .4961 ± .0310 .4781 ± .0337 .4508 ± .0290
30 .4066 ± .0409 .3761 ± .0310 .3488 ± .0400 .3373 ± .0300
40 .3351 ± .0480 .3110 ± .0383 .2686 ± .0456 .2762 ± .0358
50 .2758 ± .0480 .2626 ± .0404 .2343 ± .0434 .2293 ± .0370
13-16 10 .6506 ± .0297 .6310 ± .0272 .6273 ± .0261 .5993 ± .0253
20 .5240 ± .0280 .4959 ± .0262 .4963 ± .0266 .4580 ± .0245
30 .4474 ± .0303 .4115 ± .0262 .4089 ± .0275 .3692 ± .0243
40 .3876 ± .0340 .3487 ± .0268 .3408 ± .0319 .3071 ± .0252
50 .3330 ± .0381 .2981 ± .0281 .2764 ± .0347 .2607 ± .0263
60 .2886 ± .0434 .2598 ± .0317 .2313 ± .0373 .2255 ± .0287

Yuan et al.

Yuan et al. BMC Bioinformatics 2012 13:107   doi:10.1186/1471-2105-13-107

Open Data