Skip to main content

Table 6 The performance of multi-class relational classifier across three different SemRep datasets

From: Formalizing biomedical concepts from textual definitions

  Top 5 relations Top 10 relations All relations
F-measure    
(with Types) 94% 89.1% 82.7%
F-measure    
(only Types) 93.5% 79.2% 65.5%
Size 860 (63%) 1,144 (84%) 1,357 (100%)
  1. The size of each dataset is specified by the absolute number of instances and by the percentage of instances covered by the respective set of relations. The table reports F-Measure for two settings: including semantic types in the feature space, and excluding them.