Shallow context analysis for German idiom detection

In order to differentiate between figurative and literal usage of verb-noun combinations for the shared task on the disambiguation of German Verbal Idioms issued for KONVENS 2021, we apply and extend an approach originally developed for detecting idioms in a dataset consisting of random ngram samples. The classification is done by implementing a rather shallow, statistics-based pipeline without intensive preprocessing and examinations on the morphosyntactic and semantic level. We describe the overall approach, the differences between the original dataset and the dataset of the KONVENS task, provide experimental classification results, and analyse the individual contributions of our feature sets.

Metadaten
Author:	Miriam Amin ORCiD GND, Peter Fankhauser ORCiD GND, Marc Kupietz ORCiD GND, Roman Schneider ORCiD GND
URN:	urn:nbn:de:bsz:mh39-119560
DOI:	https://doi.org/10.5281/zenodo.5769519
Parent Title (English):	Proceedings of the shared task on the disambiguation of German verbal idioms at KONVENS 2021, Düsseldorf, Germany
Publisher:	Zenodo
Place of publication:	Genf
Document Type:	Conference Proceeding
Language:	English
Year of first Publication:	2021
Date of Publication (online):	2023/06/16
Publishing Institution:	Leibniz-Institut für Deutsche Sprache (IDS)
Publicationstate:	Veröffentlichungsversion
Reviewstate:	Peer-Review
Tag:	Natural Language Processing; idiom detection; multiword expressions; shared task
GND Keyword:	Automatische Sprachanalyse; Computerlinguistik; Datensatz; Deutsch; Kontextanalyse; Phraseologie
Page Number:	9
DDC classes:	400 Sprache / 400 Sprache, Linguistik
Open Access?:	ja
Leibniz-Classification:	Sprache, Linguistik
Linguistics-Classification:	Computerlinguistik
Program areas:	G2: Sprachinformationssysteme
Program areas:	S1: Korpuslinguistik
Program areas:	S2: Forschungskoordination und –infrastrukturen
Licence (English):	Creative Commons - Attribution 4.0 International

Open Access