A harmonised testsuite for POS tagging of German social media data

• Verfasst von: Rehbein, Ines [VerfasserIn]
• Verfasst von: Ruppenhofer, Josef [VerfasserIn]
• Verfasst von: Zimmermann, Victor [VerfasserIn]
• Titel: A harmonised testsuite for POS tagging of German social media data
• Verf.angabe: Ines Rehbein, Josef Ruppenhofer, Victor Zimmermann
• E-Jahr: 2018
• Jahr: 29.09.2018
• Umfang: 11 S.
• Fussnoten: Gesehen am 04.05.2020
• Titel Quelle: Enthalten in: The 27th International Conference on Computational Linguistics - proceedings of the conference
• Ort Quelle: [Stroudsburg, PA] : Association for Computational Linguistics, 2018
• Jahr Quelle: 2018
• Band/Heft Quelle: (2018), Seite 18-28
• ISBN Quelle: 978-1-948087-50-6
• Datenträger: Online-Ressource
• Sprache: ger
• Bibliogr. Hinweis: Forschungsdaten: Rehbein, Ines: A harmonised testsuite for social media POS tagging (DE)
• Bibliogr. Hinweis: Forschungsdaten: Rehbein, Ines: Pre-trained POS tagging models for German social media
• K10plus-PPN: 1697045855

Abstract

We present a testsuite for POS tagging German web data. Our testsuite provides the original raw text as well as the gold tokenisations and is annotated for parts-of-speech. The testsuite includes a new dataset for German tweets, with a current size of 3,940 tokens. To increase the size of the data, we harmonised the annotations in already existing web corpora, based on the Stuttgart-Tübingen Tag Set. The current version of the corpus has an overall size of 48,344 tokens of web data, around half of it from Twitter. We also present experiments, showing how different experimental setups (training set size, additional out-of-domain training data, self-training) influence the accuracy of the taggers. All resources and models will be made publicly available to the research community.