| Title: | Automatic punctuation annotation in czech broadcast news speech |
| Other Titles: | Automatická anotace interpunkce v řečových nahrávkách českých zpráv |
| Authors: | Kolář, Jáchym Švec, Jan Psutka, Josef |
| Citation: | KOLÁŘ, Jáchym; ŠVEC, Jan; PSUTKA, Josef. Automatic punctuation annotation in czech broadcast news speech. In: SPECOM 2004 Proceedings. St. Petersburg: Institute for Informatics and Automation of RAS (SPIIRAS), 2004, p. 319-325. ISBN 5-7452-0110-X. |
| Issue Date: | 2004 |
| Publisher: | SPIIRAS |
| Document type: | článek article |
| URI: | http://www.kky.zcu.cz/cs/publications/KolarJ_2004_Automaticpunctuation http://hdl.handle.net/11025/17116 |
| ISBN: | 5-7452-0110-X |
| Keywords: | automatická interpunkce;prozodie;hranice vět;rozhlasové zprávy;morfologické značkování |
| Keywords in different language: | automatic punctuation;prosody;sentence boundary;broadcast news;tag-based models |
| Abstract: | Tento článek se zabývá našimi počátečními experimenty s automatickou anotací interpunkce v mluvené češtině. Použili jsme 2 statistické modely - prozodický a jazykový. Byly otestovány 2 implementace prozodického modelu - CART a MLP. Pro jazykové modelováni byl použit N-gramový model se skrytými událostmi. Kombinovaný model dosáhl na referenčních přepisech přesnosti 95.2% a F-measure 78.2%. |
| Abstract in different language: | This paper reports our initial experiments with automatic punctuation annotation from speech. We have focused on Czech broadcast news speech. We employed two statistical models - prosodic model and language model. The prosodic model expresses relationships between prosodic quantities (such as pitch, speaking rate or loudness) and punctuation marks. We tested two implementations of this model -- decision tree and multi-layer perceptron. Hidden-event N-gram models were employed for language modeling. Instead of using an ordinary word-based model, we replaced infrequent word forms by their morphological tags and trained a mixed model. Scores from both models can be combined. The model combining language model with the decision tree yielded superior results. Testing on true words we achieved classification accuracy 95.2% and F-measure 78.2%. |
| Rights: | © Jáchym Kolář - Jan Švec - Josef Psutka |
| Appears in Collections: | Články / Articles (KKY) |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| KolarJ_2004_Automaticpunctuation.pdf | Plný text | 94,95 kB | Adobe PDF | View/Open |
Please use this identifier to cite or link to this item:
http://hdl.handle.net/11025/17116Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.