DMCA
Compound decomposition in Dutch large vocabulary speech recognition (2003)
| Venue: | IN PROCEEDINGS OF EUROSPEECH 2003, GENEVE |
| Citations: | 13 - 0 self |
BibTeX
@INPROCEEDINGS{Ordelman03compounddecomposition,
author = {Roeland Ordelman and Arjan van Hessen and Franciska de Jong},
title = {Compound decomposition in Dutch large vocabulary speech recognition},
booktitle = {IN PROCEEDINGS OF EUROSPEECH 2003, GENEVE},
year = {2003},
pages = {225--228},
publisher = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
This paper addresses compound splitting for Dutch in the context of broadcast news transcription. Language models were created using original text versions and text versions that were decomposed using a data-driven compound splitting algorithm. Language model performances were compared in terms of outof -vocabulary rates and word error rates in a real-world broadcast news transcription task. It was concluded that compound splitting does improve ASR performance. Best results were obtained when frequent compounds were not decomposed.
Keyphrases
dutch large vocabulary speech recognition compound decomposition text version asr performance real-world broadcast news transcription task language model model performance data-driven compound splitting algorithm compound splitting outof vocabulary rate frequent compound broadcast news transcription original text version word error rate
