Naslov
GDEX for Japanese: Automatic extraction of good dictionary examples

Autori
Srdanović, Irena ; Kosem, Iztok

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, kratko priopćenje, ostalo

Izvornik
GLOBALEX 2016 Lexicographic Resources for Human Language Technology / Kernerman, Illan ; Kosem, Iztok ; Krek, Simon ; Trap-Jensen, Lars - , 2016, 57-64

Skup
GLOBALEX 2016 Lexicographic Resources for Human Language Technology

Mjesto i datum
Portorož, Slovenija, 24.05.2016

Vrsta sudjelovanja
Radionica

Vrsta recenzije
Podatak o recenziji nije dostupan

Ključne riječi
dictionary examples ; corpus lexicography ; Japanese language ; difficulty levels ; automatic extraction

Sažetak
The GDEX tool, devised to assist lexicographers in identifying good dictionary examples, was initially created for the English language (Kilgarriff et al., 2008) and proved very useful in various dictionary projects (c.f. Rundell & Kilgarriff, 2011). Later on, GDEX configurations were developed for Slovene (Kosem et al., 2011, 2013) and other languages. This paper employs similar methods to design GDEX for Japanese in order to extract good example candidates from Japanese language corpora available inside the Sketch Engine system. Criteria and parameters, which were adapted to Japanese language needs, were based on the configuration for Slovene as well as the default language independent configuration available in the Sketch Engine. A number of different configurations were devised and compared in order to identify optimal values for good example identification. The paper also explores a language-learner oriented approach to good example extraction by taking into account different difficulty levels of lexemes based on the Japanese Language Proficiency Test list of words and levels. For this purposes, additional configurations were devised, which are tailored to individual levels and thus useful for language learners and lexicographers of learner’s dictionaries.

Izvorni jezik
Engleski

POVEZANOST RADA