Assessing agreement on classification tasks: the kappa statistic
Computational Linguistics
Automatic labeling of semantic roles
Computational Linguistics
Semantic Classes in Czech Valency Lexicon
TSD '08 Proceedings of the 11th international conference on Text, Speech and Dialogue
Hi-index | 0.00 |
We introduce a project to enhance a valency lexicon of Czech verbs with semantic roles. For this purpose, we make use of FrameNet. At the present stage, frame elements from FrameNet have been mapped to valency complementations of verbs of communication and verbs of exchange. The feasibility of this task has been proven by the achieved inter-annotator agreement --- 95.6% for the verbs of communication and 91.2% for the verbs of exchange. As a result, we have obtained 37 semantic roles for the verbs of communication and 34 for the verbs of exchange, based on frame elements of upper level semantic frames from FrameNet.