Skip to main content

Table 5 Results of using different NLP toolkits on the BC5CDR-chemical dataset

From: Improving biomedical named entity recognition with syntactic information

 

BioBERT-base

BioBERT-large

 

F1

\(\sigma\)

F1

\(\sigma\)

Baseline

93.50

0.10

93.90

0.31

 Stanford CoreNLP Toolkits

    

  PL (\({\mathcal {M}}\))

93.73

0.19

94.05

0.23

  DR (\({\mathcal {M}}\))

\(\mathit{93} .\mathit{78}\)

0.18

94.05

0.10

 spaCy

    

  PL (\({\mathcal {M}}\))

93.69

0.12

\(\mathit{94} .\mathit{06}\)

0.10

  DR (\({\mathcal {M}}\))

93.71

0.12

93.97

0.13

  1. The experimental results [the average F1 scores and the standard deviation (\(\sigma\))] of our method with KVMN (\({\mathcal {M}}\)) using different NLP toolkits (i.e., Stanford CoreNLP Toolkits and spaCy) to obtain POS labels (PL) and dependency relations (DR). The results of baseline methods without using any syntactic information are also reported for reference