SBLC: a hybrid model for disease named entity recognition based on semantic bidirectional LSTMs and conditional random fields

Table 7 Effects of different parameter settings and the final optimized result

Parameter	Precision	Recall	F1
CRF	0.701	0.675	0.688
Bi-LSTM	0.600	0.425	0.498
Ab3p + CRF	0.726	0.689	0.707
Ab3p + Bi-LSTM	0.645	0.452	0.532
Bi-LSTM + CRF	0.806	0.800	0.803
Ab3p + Bi-LSTM + CRF	0.813	0.808	0.811
Word Embedding + Bi-LSTM	0.675	0.501	0.575
Word Embedding + CRF	0.821	0.772	0.796
Word Embedding + Bi-LSTM + CRF	0.842	0.828	0.835
Ab3p + Word Embedding + Bi-LSTM	0.613	0.689	0.648
Ab3p + Word Embedding + CRF	0.846	0.786	0.815
Ab3p + Word Embedding + Bi-LSTM + CRF (SBLC)	0.866	0.858	0.862

ISSN: 1472-6947