Skip to main content

Table 12 Comparison of semantic role induction results with those of previous approaches in terms of performance on the test data

From: Semantic frame induction through the detection of communities of verbs and their arguments

  Purity F1 BCubed F1
Baseline 56.05 39.03
Ribeiro et al. (2019) 64.10 45.66
Anwar et al. (2019) 62.00 42.10
Arefyev et al. (2019) 77.47 64.04
Our Approach (Dev. Threshold) 64.96 48.85
  1. The results of (Arefyev et al. 2019) are crossed because they were obtained using a supervised approach