Skip to Main Content
Table 10

The frequency and P / R / F1 scores for typical argument role labels on the CoNLL-2009 English test set. Baseline refers to the tree-based model, and +Pruning refers to the SynRule Soft Pruning method. Due to the width constraint, we omit the Precision (P) and Recall (R) scores of the nominal pred.

verbal prednominal pred
RoleFREQBaseline+PruningFREQBaseline+Pruning
A0 15% 93.1 / 91.9 / 92.5 93.2 / 92.3 / 92.7 10% 83.3 83.6 
A1 21% 93.9 / 93.1 / 93.5 93.6 / 93.4 / 93.5 16% 87.2 87.4 
A2 5% 84.3 / 81.8 / 83.0 85.1 / 83.1 / 84.1 7% 81.0 82.8 
AM-* 16% 82.2 / 80.2 / 81.2 82.4 / 80.6 / 81.5 5% 75.4 76.3 
verbal prednominal pred
RoleFREQBaseline+PruningFREQBaseline+Pruning
A0 15% 93.1 / 91.9 / 92.5 93.2 / 92.3 / 92.7 10% 83.3 83.6 
A1 21% 93.9 / 93.1 / 93.5 93.6 / 93.4 / 93.5 16% 87.2 87.4 
A2 5% 84.3 / 81.8 / 83.0 85.1 / 83.1 / 84.1 7% 81.0 82.8 
AM-* 16% 82.2 / 80.2 / 81.2 82.4 / 80.6 / 81.5 5% 75.4 76.3 
Close Modal

or Create an Account

Close Modal
Close Modal