Skip to main content

Table 4 Effect of the weighting and covariate shift methods on the development data sets.

From: Adaptable, high recall, event extraction system with minimal configuration

Task Weighting Covariate shift Recall Precision F-score (%)
CG × × 40.70 62.19 49.20
  × 50.11 50.68 50.39
  × 44.42 59.78 50.96
  48.83 54.10 51.33
PC × × 37.89 61.26 46.82
  × 44.23 49.18 46.57
  × 40.65 55.81 47.04
  42.73 52.13 46.97