Skip to main content

Table 8 Performance comparison (maximum F score) with existing models on the DDI corpus

From: Relation extraction between bacteria and biotopes from biomedical texts with attention mechanisms and domain-specific contextual representations

SVM-basedUturku [51]59.449.973.2
 FBK-irst [52]65.165.664.6
 RAIHANI [53]71.168.773.7
DL-basedLiu-CNN [9]69.864.775.7
 MCCNN [54]70.265.376.0
 TEES-CNN [55]73.567.680.5
 Joint AB-LSTM [13]71.565.074.5
 Char-RNNs [42]72.165.980.0
 Hierarchy RNN [14]72.971.874.1
 Recursive NN [56]73.569.677.8
 Attn-BLSTM [30]77.376.878.4
 Our model80.383.077.6
  1. The listed results derive from the corresponding papers. F: F score (micro); R: recall; P: precision. The highest scores are highlighted in bold