Skip to main content

Advertisement

Table 3 Stacking experiments for the EPI track

From: Combining joint models for biomedical event extraction

System Recall Precision F1
UMass 56.7 73.2 63.9
Stanford (1N) 52.2 79.0 62.9
Stanford (1P) 51.7 78.2 62.3
Stanford (2N) 48.1 82.6 60.8
Stanford (2P) 51.9 77.6 62.2
Stanford (1N, reranked) 57.7 73.2 64.6
Stanford (1P, reranked) 57.6 70.4 63.3
Stanford (2N, reranked) 55.3 75.4 63.8
Stanford (2P, reranked) 56.9 71.3 63.3
Stanford (1N+2P, reranked) 57.9 71.1 63.8
Stanford (all, reranked) 57.0 73.1 64.1
UMassā†Stanford (all) (= FAUST) 57.9 79.7 67.1
  1. BioNLP F1 scores on the development set of EPI using the CORE metric.