Skip to main content

Advertisement

Table 8 Detailed performance results on highly related domains with different training modes, including the Basic Model A (trained on the training and development sets of DataMLEE), the Basic Model A (trained on the combination of DataST09 and the training and development sets of DataMLEE), and our MLTrigNer Model (using DataMLEE as the target dataset and DataST09 as the source dataset)

From: Multiple-level biomedical event trigger recognition with transfer learning

Trigger Recognition System Precision Recall F1-Measure
Basic Model A (DataMLEE) 81.63 74.26 77.77
Basic Model A (DataMLEE + DataST09) 78.78 73.92 76.28
Our MLTrigNer Model (DataMLEE + DataST09) 83.31 79.40 81.31
  1. In these results, the best F1 value of our MLTrigNer model is marked in bold