Abstract
This paper discusses elliptical compounds, which are frequently used in German in order to avoid repetitions. This phenomenon involves truncated words, mostly truncated compounds. These words pose a challenge in PoS tagging and lemmatization, which often leads to unknown or incomplete lemmas. We present an approach to reconstruct complete lemmas of truncated compounds in order to improve subsequent language technology or corpus linguistic applications. Results show an f-measure of 95.6% for the detection of elliptical compound patterns and 86.4% for the correction of compound lemmas. © 2013 Springer-Verlag.
Author supplied keywords
Cite
CITATION STYLE
Aepli, N., & Volk, M. (2013). Reconstructing complete lemmas for incomplete German compounds. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8105 LNAI, pp. 1–13). https://doi.org/10.1007/978-3-642-40722-2_1
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
