Volltext-Downloads (blau) und Frontdoor-Views (grau)

Data point selection for self-training

  • Problems for parsing morphologically rich languages are, amongst others, caused by the higher variability in structure due to less rigid word order constraints and by the higher number of different lexical forms. Both properties can result in sparse data problems for statistical parsing. We present a simple approach for addressing these issues. Our approach makes use of self-training on instances selected with regard to their similarity to the annotated data. Our similarity measure is based on the perplexity of part-of-speech trigrams of new instances measured against the annotated training data. Preliminary results show that our method outperforms a self-training setting where instances are simply selected by order of occurrence in the corpus and argue that selftraining is a cheap and effective method for improving parsing accuracy for morphologically rich languages.

Export metadata

Additional Services

Share in Twitter Search Google Scholar

Statistics

frontdoor_oas
Metadaten
Author:Ines Rehbein
URN:urn:nbn:de:bsz:mh39-56043
URL:https://aclanthology.info/pdf/W/W11/W11-3800.pdf
ISBN:978-1-932432-73-2
Parent Title (English):Proceedings of SPMRL 2011. The Second Workshop on Statistical Parsing of Morphologically Rich Languages (SPMRL 2011). October 6, 2011. Dublin, Ireland
Publisher:Association for Computational
Place of publication:Stroudsburg, PA
Document Type:Conference Proceeding
Language:English
Year of first Publication:2011
Date of Publication (online):2016/11/21
GND Keyword:Automatische Sprachanalyse; Satzanalyse
First Page:62
Last Page:67
Dewey Decimal Classification:400 Sprache / 400 Sprache, Linguistik
Linguistics-Classification:Computerlinguistik
Open Access?:Ja
Licence (English):License LogoCreative Commons - Attribution-NonCommercial-ShareAlike 3.0 Unported