An Improved Bulgarian Natural Language Processing Pipeline

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F23%3A7JMFCF9C" target="_blank" >RIV/00216208:11320/23:7JMFCF9C - isvavai.cz</a>

Výsledek na webu

<a href="https://www.researchgate.net/profile/Melania-Berbatova/publication/371081880_An_Improved_Bulgarian_Natural_Language_Processing_Pipeline/links/64787b68b3dfd73b7758815e/An-Improved-Bulgarian-Natural-Language-Processing-Pipeline.pdf" target="_blank" >https://www.researchgate.net/profile/Melania-Berbatova/publication/371081880_An_Improved_Bulgarian_Natural_Language_Processing_Pipeline/links/64787b68b3dfd73b7758815e/An-Improved-Bulgarian-Natural-Language-Processing-Pipeline.pdf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.60063/gsu.fmi.110.37-50" target="_blank" >10.60063/gsu.fmi.110.37-50</a>

Jazyk výsledku

angličtina

Název v původním jazyce

An Improved Bulgarian Natural Language Processing Pipeline

Popis výsledku v původním jazyce

"In this paper, we present a language pipeline for processing Bulgarian language data. The pipeline consistsnof the following steps: tokenization, sentence splitting, part-of-speech tagging, dependency parsing,nnamed entity recognition, lemmatization, and word sense disambiguation. The rst two components arenbased on rules and lists of words specic to the Bulgarian language, while the rest of the components usenmachine learning algorithms trained on universal dependency data and pretrained word vectors. Thenpipeline is implemented in the Python library spaCy and achieves signicant results on all the includednsubtasks. The pipeline is open source and is available on Github for use by researchers and developersnfor a variety of natural language processing and text analysis tasks."

Název v anglickém jazyce

An Improved Bulgarian Natural Language Processing Pipeline

Popis výsledku anglicky

"In this paper, we present a language pipeline for processing Bulgarian language data. The pipeline consistsnof the following steps: tokenization, sentence splitting, part-of-speech tagging, dependency parsing,nnamed entity recognition, lemmatization, and word sense disambiguation. The rst two components arenbased on rules and lists of words specic to the Bulgarian language, while the rest of the components usenmachine learning algorithms trained on universal dependency data and pretrained word vectors. Thenpipeline is implemented in the Python library spaCy and achieves signicant results on all the includednsubtasks. The pipeline is open source and is available on Github for use by researchers and developersnfor a variety of natural language processing and text analysis tasks."

Druh

J_ost - Ostatní články v recenzovaných periodicích

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

—

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

"ANNUAL OF SOFIA UNIVERSITY “ST. KLIMENT OHRIDSKI”FACULTY OF MATHEMATICS AND INFORMATICS"

ISSN

1313-9215

e-ISSN

—

Svazek periodika

110

Číslo periodika v rámci svazku

2023

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

14

Strana od-do

37-50

Kód UT WoS článku

—

EID výsledku v databázi Scopus

—