Automated Semantic Classification of French Verbs
Abstract
The aim of this work is to explore (semi-)automatic means to create
a Levin-type classification of French verbs, suitable for Natural
Language Processing. For English, a classification based on Levin's
method is VerbNet (Kipper 2005). VerbNet is an extensive digital
verb lexicon which systematically extends Levin's
classes while ensuring that class members have a common semantics
and share a common set of syntactic frames and thematic roles.
In this work we reorganise the verbs from three French syntax
lexicons, namely Volem, the Grammar-Lexicon
(Ladl) and Dicovalence, into VerbNet-like verb classes using the technique of Formal Concept Analysis.
We automatically acquire syntactic-semantic verb class and diathesis
alternation information. We create large scale verb classes and
compare their verb and frame
distributions to those of VerbNet.
We discuss possible
evaluation schemes and finally focus on an evaluation methodology
with respect to VerbNet, of which we present the theoretical motivation and
analyse the feasibility on a small hand-built example.
Origin : Files produced by the author(s)
Loading...