Paper: Benchmarking of Statistical Dependency Parsers for French

ACL ID C10-2013
Title Benchmarking of Statistical Dependency Parsers for French
Venue International Conference on Computational Linguistics
Session Poster Session
Year 2010

We compare the performance of three statistical parsing architectures on the problem of deriving typed dependency structures for French. The architectures are based on PCFGs with latent vari- ables, graph-based dependency parsing and transition-based dependency parsing, respectively. We also study the inu- ence of three types of lexical informa- tion: lemmas, morphological features, and word clusters. The results show that all three systems achieve competitive per- formance, with a best labeled attachment score over 88%. All three parsers benet fromtheuseofautomaticallyderivedlem- mas, while morphological features seem to be less important. Word clusters have a positive effect primarily on the latent vari- able parser.