Titel: |
Fine-grained POS tagging of German Twitter data |
Personen: | Rehbein, Ines |
Jahr: |
2013 |
Typ: |
Aufsatz |
Verlag: |
Springer |
Ortsangabe: |
Heidelberg/Berlin |
In: |
Gurevych, Iryna/Biemann, Chris/Zesch, Torsten: Language Processing and Knowledge in the Web. Proceedings of the 25th International Conference, GSCL 2013, Darmstadt, Germany, 25 - 27 September 2013 |
Seiten: |
162-175 |
Untersuchte Sprachen: |
Deutsch*German |
Schlagwörter: |
Datenmodellierung*data modelling
mobile Endgeräte*mobile devices
Nutzerbeteiligung*user contribution
|
Abstract: |
This paper presents the first work on POS tagging German Twitter data, showing that despite the noisy and often cryptic nature of the data a fine-grained analysis of POS tags on Twitter microtext is feasible. Our CRF-based tagger achieves an accuracy of around 89% when trained on LDA word clusters, features from an automatically created dictionary and additional out-of-domain training data. |