Key points are not available for this paper at this time.
This paper presents SemFrame, a system that automatically induces the names and internal structures of semantic frames. After SemFrame identifies sets of frame-evoking verb synsets, the conceptual density of nodes in the WordNet network for corresponding nouns and noun synsets is computed and analyzed. Conceptually dense nodes are candidates for frame names and frame slots. Ca. 76% of the frame names and 87% of the frame slots generated by SemFrame are rated adequate by human judges.
Green et al. (Thu,) studied this question.