In-depth exploitation of noun and verb semantics to identify causation in verb-noun pairs

Abstract

Recognition of causality is important to achieve natural language discourse understanding. Previous approaches rely on shallow linguistic features. In this work, we propose to identify causality in verbnoun pairs by exploiting deeper semantics of nouns and verbs. Particularly, we acquire and employ three novel types of knowledge: (1) semantic classes of nouns with a high and low tendency to encode causality along with information regarding metonymies, (2) data-driven semantic classes of verbal events with the least tendency to encode causality, and (3) tendencies of verb frames to encode causality. Using these knowledge sources, we achieve around 15% improvement in Fscore over a supervised classifier trained using linguistic features.