Please use this identifier to cite or link to this item:
Type of publication: research article
Type of publication (PDB): Straipsnis recenzuojamoje užsienio tarptautinės konferencijos medžiagoje / Article in peer-reviewed foreign international conference proceedings (P1d)
Field of Science: Filologija / Philology (H004)
Author(s): Boizou, Loic;Kovalevskaitė, Jolanta;Rimkutė, Erika
Title: Automatic lemmatisation of Lithuanian MWEs
Is part of: NODALIDA 2015 : proceedings of the 20th Nordic conference of computational linguistics, May 11–13, 2015, Institute of the Lithuanian language, Vilnius / editor Beata Megyesi. Linköping : Linköping University Electronic Press, 2015
Extent: p. 41-49
Date: 2015
Series/Report no.: (NEALT proceedings, vol. 23 1650-3638)
Note: eISSN 1650-3740
Keywords: Grammatical restrictions;Lithuanian language;Lemmatisation
ISBN: 9789175190983
Abstract: This article presents a study of lemmatisation of flexible multiword expressions in Lithuanian. An approach based on syntactic analysis designed for multiword term lemmatisation was adapted for a broader range of MWEs taken from the Dictionary of Lithuanian Nominal Phrases. In the present analysis, the main lemmatisation errors are identified and some improvements are proposed. It shows that automatic lemmatisation can be improved by taking into account the whole set of grammatical forms for each MWE. It would allow selecting the optimal grammatical form for lemmatisation and identifying some grammatical restrictions
Affiliation(s): Lituanistikos katedra
Užsienio kalbų, lit. ir vert. s. katedra
Vytauto Didžiojo universitetas
Appears in Collections:Universiteto mokslo publikacijos / University Research Publications

Files in This Item:
marc.xml7.94 kBXMLView/Open

MARC21 XML metadata

Show full item record
Export via OAI-PMH Interface in XML Formats
Export to Other Non-XML Formats

CORE Recommender

Page view(s)

checked on Mar 30, 2021


checked on Mar 31, 2021

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.