Pregled bibliografske jedinice broj: 634488
Domain Dependence of Statistical Named Entity Recognition and Classification in Croatian Texts
Domain Dependence of Statistical Named Entity Recognition and Classification in Croatian Texts // Proceedings of the 35th International Conference on Information Technology Interfaces (ITI 2013) / Lužar-Stiffler, Vesna ; Jarec, Iva (ur.).
Zagreb: Sveučilišni računski centar Sveučilišta u Zagrebu (Srce), 2013. str. 277-283 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 634488 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Domain Dependence of Statistical Named Entity Recognition and Classification in Croatian Texts
Autori
Agić, Željko ; Bekavac, Božo
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Proceedings of the 35th International Conference on Information Technology Interfaces (ITI 2013)
/ Lužar-Stiffler, Vesna ; Jarec, Iva - Zagreb : Sveučilišni računski centar Sveučilišta u Zagrebu (Srce), 2013, 277-283
ISBN
978-953-7138-30-1
Skup
35th International Conference on Information Technology Interfaces (ITI 2013)
Mjesto i datum
Cavtat, Hrvatska, 24.06.2013. - 27.06.2013
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
text domain; domain dependence; named entity recognition; Croatian language
Sažetak
Influence of text domain selection on statistical named entity recognition and classification in Croatian texts is investigated. Two datasets of Croatian newspaper texts of differing text domains were manually annotated for named entities and used for training and testing the Stanford NER system for named entity recognition based on sequence labeling with CRF. State of the art scores were observed in both domains. A strong preference for systems trained on mixed text domains is established by the experiment. The top- performing system was recorded with an overall F1- score of 0.876 on mixed-domain test sets, scoring 0.899 in one of the selected domains and 0.852 in the other. The single best domain F1-scores were recorded at 0.910 and 0.858.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo, Informacijske i komunikacijske znanosti, Filologija
POVEZANOST RADA
Projekti:
130-1300646-0645 - Hrvatski jezični resursi i njihovo obilježavanje (Tadić, Marko, MZOS ) ( CroRIS)
130-1300646-1002 - Leksička semantika u izradi Hrvatskog WordNeta (Raffaelli, Ida, MZOS ) ( CroRIS)
130-1300646-1776 - Računalna sintaksa hrvatskoga jezika (Dovedan Han, Zdravko, MZOS ) ( CroRIS)
Ustanove:
Filozofski fakultet, Zagreb
