Skip to main content

Table 7 Results of ablation study with Span F1-score and Type accuracy metrics

From: Combining data augmentation and domain information with TENER model for Clinical Event Detection

Model

Span F1-score

Type accuracy

Ours

90.33

93.00

BioBERT \(\rightarrow\) Glove 100d

79.08

86.60

-data augmentation

90.01

93.05

TENER \(\rightarrow\) CNN

90.28

93.10

TENER \(\rightarrow\) BiLSTM

90.13

92.98

  1. “-” means remove component, “\(\rightarrow\)” means replace component, “TENER \(\rightarrow\) CNN” means character-level embedding component. “TENER \(\rightarrow\) BiLSTM” means encoding component