Automatická anotace angličtiny na tektogramatické rovině

Tectogrammatical layer is very complex and its annotation is di cult and expensive. Unlike other corpora, the Prague English Dependency Treebank (pedt) is based on data for which there already exists a syntactic annotation, even though a fundamentally di erent one. The goal of this work is to propos...

Full description

Bibliographic Details
Main Author: Toman, Josef
Other Authors: Hajič, Jan
Format: Dissertation
Language:Czech
Published: 2009
Online Access:http://www.nusl.cz/ntk/nusl-275611
id ndltd-nusl.cz-oai-invenio.nusl.cz-275611
record_format oai_dc
spelling ndltd-nusl.cz-oai-invenio.nusl.cz-2756112017-06-27T04:39:46Z Automatická anotace angličtiny na tektogramatické rovině Automatic annotation of English on the tectogrammatical level Toman, Josef Hajič, Jan Žabokrtský, Zdeněk Tectogrammatical layer is very complex and its annotation is di cult and expensive. Unlike other corpora, the Prague English Dependency Treebank (pedt) is based on data for which there already exists a syntactic annotation, even though a fundamentally di erent one. The goal of this work is to propose and implement methods of automatic annotation that are using the available data and (preferably) would lead to minimization of the e ort needed for a manual annotation. A high-quality evaluation is important so that the contribution of the used methods can be veri ed. Tens of modules, which focus on various aspects of annotation, were created. The analysis of their activity is complicated and required a complex system to be created. The analyses created with it are very detailed. The outcome is positive and urges to continue the work and extend it further. 2009 info:eu-repo/semantics/masterThesis http://www.nusl.cz/ntk/nusl-275611 cze info:eu-repo/semantics/restrictedAccess
collection NDLTD
language Czech
format Dissertation
sources NDLTD
description Tectogrammatical layer is very complex and its annotation is di cult and expensive. Unlike other corpora, the Prague English Dependency Treebank (pedt) is based on data for which there already exists a syntactic annotation, even though a fundamentally di erent one. The goal of this work is to propose and implement methods of automatic annotation that are using the available data and (preferably) would lead to minimization of the e ort needed for a manual annotation. A high-quality evaluation is important so that the contribution of the used methods can be veri ed. Tens of modules, which focus on various aspects of annotation, were created. The analysis of their activity is complicated and required a complex system to be created. The analyses created with it are very detailed. The outcome is positive and urges to continue the work and extend it further.
author2 Hajič, Jan
author_facet Hajič, Jan
Toman, Josef
author Toman, Josef
spellingShingle Toman, Josef
Automatická anotace angličtiny na tektogramatické rovině
author_sort Toman, Josef
title Automatická anotace angličtiny na tektogramatické rovině
title_short Automatická anotace angličtiny na tektogramatické rovině
title_full Automatická anotace angličtiny na tektogramatické rovině
title_fullStr Automatická anotace angličtiny na tektogramatické rovině
title_full_unstemmed Automatická anotace angličtiny na tektogramatické rovině
title_sort automatická anotace angličtiny na tektogramatické rovině
publishDate 2009
url http://www.nusl.cz/ntk/nusl-275611
work_keys_str_mv AT tomanjosef automatickaanotaceanglictinynatektogramatickerovine
AT tomanjosef automaticannotationofenglishonthetectogrammaticallevel
_version_ 1718467938791456768