Universal Dependencies for the AnCora treebanks

Abstract : Abstract: The present article describes the conversion of the Catalan and Spanish AnCora treebanks to the Universal Dependencies formalism. We describe the conversion process and assess the quality of the resulting treebank in terms of parsing accuracy by means of monolingual, cross-lingual and cross-domain parsing evaluation. The converted treebanks show an internal consistency comparable to the one shown by the original CoNLL09 distribution of AnCora, and indicate some differences in terms of multiword expression inventory with regards to the already existing UD Spanish treebank. The two new converted treebanks will be released in version 1.3 of Universal Dependencies.
Complete list of metadatas

Cited literature [8 references]  Display  Hide  Download

https://hal.inria.fr/hal-01426751
Contributor : Héctor Martínez Alonso <>
Submitted on : Wednesday, January 4, 2017 - 7:21:45 PM
Last modification on : Friday, January 4, 2019 - 5:33:38 PM
Long-term archiving on : Wednesday, April 5, 2017 - 3:25:46 PM

File

5341-4677-1-PB.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01426751, version 1

Collections

Citation

Hector Martinez Alonso, Daniel Zeman. Universal Dependencies for the AnCora treebanks . Procesamiento del Lenguaje Natural, Sociedad Espanola para el Procesamiento del Lenguaje Natural, 2016, ⟨http://journal.sepln.org/sepln/ojs/ojs/index.php/pln/issue/view/220⟩. ⟨hal-01426751⟩

Share

Metrics

Record views

170

Files downloads

110