Exploiting and integrating rich features for biological literature classification

BMC Bioinformatics

Table 9 Classifier-level Integration.

*Feature*	*Precision*	*Recall*	*F-Score*	*AUC*
Unigram + Bigram	0.7015	0.8213	0.7567	0.8036
Output based Integration	0.7248	0.8853	0.7971	0.8539
AdaBoost	0.7995	0.8933	0.8438	0.8746

(Normalize the output of each classifier and unify them into new feature vectors)

Integration on length-fixed string feature, entity feature and template feature

ISSN: 1471-2105