From Dependencies to Constituents in BDT (Basque Dependency Treebank)

keywords: 
Treebank, dependency-based, constituent-based, model turning, equivalences
Description: 
In this work a process for turning a dependency-based corpus to a constituent-based one has to be designed and developed. Some previous steps have been already carried out: the equivalence rules for basic syntactic phenomena have been proposed, evaluated and refined. Now, the aim is to study in depth other complex phenomena, such as all kind of MWLUs and composed and subordinated sentences, make an evaluation and get almost all the BDT corpus in both formalisms.
Objectives: 
To set up the equivalence rules for turning a dependency-based model into a constituent-based one, giving a solution to a great number of syntactic phenomena.
Task: 
To understand both the dependency- and constituent-based model, design equivalence rules, and make it possible to get automatically corpora analized in both formalisms.
References: 
Aldezabal I., Aranzabe M.J., Diaz de Ilarraza A., Fernández K. (2008) “From Dependencies to Constituents in the Reference Corpus for the Processing of Basque” Procesamiento del Lenguaje Natural, 41, 147-154.
Team: 
Izaskun Aldezabal, María Jesús Aranzabe, Kike Fernández
Profile: 
Linguist
contact: 
izaskun.aldezabal[abildua|at]ehu.eus
Date: 
2017