Skip to main content

Table 2 Statistics of the data sets after modification

From: A generalizable NLP framework for fast development of pattern-based biomedical relation extraction systems

Events   Training set   Development set
   (%)   (%)
Simple Event 3,165 84.92 923 80.19
 Gene_expression 2,094 86.64 614 79.23
 Transcription 511 72.59 115 69.28
 Protein_catabolism 105 92.11 22 95.65
 Phosphorylation 185 94.87 107 95.54
 Localization 270 90.91 65 86.67
Binding 874 71.00 380 75.55
Total 4,039 81.46 1,303 78.78
  1. Statistics of events with selected triggers on BioNLP-ST 2011 ST GE task. If an event’s argument is within an equivalence relation with n members, this event will be counted n times. % = Events with selected triggers/All events.