Automated Lossless Hyper-Minimization for Morphological Analyzers

Show full item record



Permalink

http://hdl.handle.net/10138/159354

Citation

Drobac , S , Silfverberg , M & Linden , K 2015 , Automated Lossless Hyper-Minimization for Morphological Analyzers . in T Hanneforth & C Wurm (eds) , Proceedings of the 12th International Conference on Finite-State Methods and Natural Language Processing 2015 : Collected papers . ACL , Stroudsburg , International Conference on Finite-State Methods and Natural Language Processing , Düsseldorf , Germany , 22/06/2015 .

Title: Automated Lossless Hyper-Minimization for Morphological Analyzers
Author: Drobac, Senka; Silfverberg, Miikka; Linden, Krister
Editor: Hanneforth, Thomas; Wurm, Christian
Contributor: University of Helsinki, Department of Modern Languages 2010-2017
University of Helsinki, Department of Modern Languages 2010-2017
University of Helsinki, Department of Modern Languages 2010-2017
Publisher: ACL
Date: 2015-04-22
Language: eng
Number of pages: 5
Belongs to series: Proceedings of the 12th International Conference on Finite-State Methods and Natural Language Processing 2015 Collected papers
ISBN: 978-1-950737-97-0
URI: http://hdl.handle.net/10138/159354
Abstract: This paper presents a fully automated lossless hyper-minimization method for finite state morphological analyzers in Xerox lexc formalism. The method utilizes flag diacritics to preserve the structure of the original lexc description in the finite-state analyzer, which results in reduced size of the analyzer. We compare our method against an earlier solution by Drobac et al. (2014) which requires manual selection of flag diacritics and results in slow lookup. We show that our method gives similar size reductions while maintaining fast lookup without requiring any manual input.
Subject: 6121 Languages
113 Computer and information sciences
Rights:


Files in this item

Total number of downloads: Loading...

Files Size Format View
W15_4806.pdf 92.22Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record