On the problems of creating a golden standard of inflected forms in Portuguese
Diana Santos (1), Anabela Barreiro (2)
(1) SINTEF ICT, Pb 124 Blindern, 0314 Oslo, Norway, Diana.Santos@sintef.no; (2) University of Lisbon and CAUTL (IST), Av. Rovisco Pais, 1049-001 Lisboa, Portugal, Anabela.Barreiro@label.ist.utl.pt
This paper describes our attempt to build a consensus round the morphological analysis of a set of forms for Portuguese, to be used as a basis for the creation of a “golden list” in the first Morpholympics for Portuguese, Morfolimpíadas, an evaluation contest on Portuguese morphological analysis. This golden standard was used to rank participating morphological systems according to precision and coverage. The discussion in the paper is centered on the general choices made and the problems encountered. The paper ends with a short description of the (publicly available) resource.
evaluation, morphology, golden standard, evaluation contests, Portuguese