Table 2 Performance and configurations of all the evaluations in the protein/gene name recognition task
From: Recognition of protein/gene names from text using an ensemble of classifiers
Modules | Closed-1 | Closed-2 | Closed-3 | Open-1 |
|---|---|---|---|---|
SVM | Surface word, orthographic feature, morphological pattern, trigger word | |||
| Â | GENIA-POS | Refined-BioCreative-POS | Refined-BioCreative-POS | Refined-BioCreative-POS |
DHMM1 | Surface word, orthographic feature | |||
| Â | GENIA-POS | Refined-BioCreative-POS | Refined- BioCreative-POS | Refined-BioCreative-POS |
DHMM2 | Surface word, orthographic feature, BioCreative-POS | |||
Ensemble | Majority Voting | |||
Abbreviation Resolution | Abbreviation Resolution based on the parentheses structure | |||
Name Refinement | N/A | N/A | YES | N/A |
Dictionary Matching | Closed Dictionary | Closed Dictionary | Closed Dictionary | Open Dictionary |
Overall Performance | P79.97 | P80.46 | P82.00 | P75.10 |
| Â | R80.50 | R80.80 | R83.17 | R81.26 |
| Â | F80.23 | F80.63(+0.40) | F82.58(+2.35) | F78.06(-4.52) |