Steiner, Petra (2019)
Combining Data-Intense and Compute-Intense Methods for Fine-Grained Morphological Analyses.
2nd International Workshop on Resources and Tools for Derivational Morphology. Prague, Czech Republic (19.-20.09.2019)
Konferenzveröffentlichung, Bibliographie
Kurzbeschreibung (Abstract)
This article describes a hybrid approach for German derivational and compositional morphology. Its first module is based on retrieval from morphological databases. The second module builds on the results of a word segmenter and uses a context-based approach by exploiting 1.8 million texts from Wikipedia for the disambiguation of multiple morphological splits. Insights from Quantitative Linguistics help countering two sparse-data problems. The results can become more fine-grained during each cycle of computation and be added to the lexical input data with or without supervision. The evaluation on an inflight magazine shows a good coverage and an accuracy of 93% for the deep-level analyses
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2019 |
| Autor(en): | Steiner, Petra |
| Art des Eintrags: | Bibliographie |
| Titel: | Combining Data-Intense and Compute-Intense Methods for Fine-Grained Morphological Analyses |
| Sprache: | Deutsch |
| Publikationsjahr: | 21 September 2019 |
| Verlag: | ACL |
| Buchtitel: | DeriMO 2019: Proceedings of the Second International Workshop on Resources and Tools for Derivational Morphology |
| Veranstaltungstitel: | 2nd International Workshop on Resources and Tools for Derivational Morphology |
| Veranstaltungsort: | Prague, Czech Republic |
| Veranstaltungsdatum: | 19.-20.09.2019 |
| URL / URN: | https://aclanthology.org/W19-8506 |
| Kurzbeschreibung (Abstract): | This article describes a hybrid approach for German derivational and compositional morphology. Its first module is based on retrieval from morphological databases. The second module builds on the results of a word segmenter and uses a context-based approach by exploiting 1.8 million texts from Wikipedia for the disambiguation of multiple morphological splits. Insights from Quantitative Linguistics help countering two sparse-data problems. The results can become more fine-grained during each cycle of computation and be added to the lexical input data with or without supervision. The evaluation on an inflight magazine shows a good coverage and an accuracy of 93% for the deep-level analyses |
| Fachbereich(e)/-gebiet(e): | Zentrale Einrichtungen Zentrale Einrichtungen > Universitäts- und Landesbibliothek (ULB) |
| Hinterlegungsdatum: | 19 Jun 2023 13:33 |
| Letzte Änderung: | 19 Jun 2023 13:33 |
| PPN: | |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum