Derivational Morphemes as Markers of Borrowed Words in Czech

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3A10511658" target="_blank" >RIV/00216208:11320/25:10511658 - isvavai.cz</a>

Výsledek na webu

<a href="https://events.unifr.ch/derimo2025/en/assets/public/files/derimo_2025_proceedings.pdf" target="_blank" >https://events.unifr.ch/derimo2025/en/assets/public/files/derimo_2025_proceedings.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Derivational Morphemes as Markers of Borrowed Words in Czech

Popis výsledku v původním jazyce

In Czech, borrowed words are often marked by the presence of morphemes of foreign origin. Non- native derivational affixes such as un-, dys-, and anti- frequently appear in these words, typically alongside foreign stems. However, it is unclear to what extent these morphemes independently influence the classification of a word as borrowed or act as a marker for borrowed words. This study investigates the role of derivational morphemes in marking borrowed words in four parts-of- speech (POS) categories: nouns, verbs, adjectives, and adverbs. We extract native and borrowed words from DeriNet, which categorizes words based on their POS tags and loanword status. Using a multinomial Naive Bayes classifier, we perform binary classification to distinguish native and borrowed words, extracting classification probabilities for the morphemes that make up these words. We compare these results with attention-based binary LSTM classifier and RobeCzech, a pre-trained RoBERTa model for Czech that we finetune for our

Název v anglickém jazyce

Derivational Morphemes as Markers of Borrowed Words in Czech

Popis výsledku anglicky

In Czech, borrowed words are often marked by the presence of morphemes of foreign origin. Non- native derivational affixes such as un-, dys-, and anti- frequently appear in these words, typically alongside foreign stems. However, it is unclear to what extent these morphemes independently influence the classification of a word as borrowed or act as a marker for borrowed words. This study investigates the role of derivational morphemes in marking borrowed words in four parts-of- speech (POS) categories: nouns, verbs, adjectives, and adverbs. We extract native and borrowed words from DeriNet, which categorizes words based on their POS tags and loanword status. Using a multinomial Naive Bayes classifier, we perform binary classification to distinguish native and borrowed words, extracting classification probabilities for the morphemes that make up these words. We compare these results with attention-based binary LSTM classifier and RobeCzech, a pre-trained RoBERTa model for Czech that we finetune for our

Druh

D - Stať ve sborníku

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

S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2025

Kód důvěrnosti údajů

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

Název statě ve sborníku

Proceedings of Fifth International Workshop on Resources and Tools for Derivational Morphology

ISBN

978-2-8399-4786-2

ISSN

—

e-ISSN

—

Počet stran výsledku

10

Strana od-do

109-118

Název nakladatele

University of Fribourg

Místo vydání

Fribourg, Switzerland

Místo konání akce

Fribourg, Switzerland

Datum konání akce

4. 9. 2025

Typ akce podle státní příslušnosti

EUR - Evropská akce

Kód UT WoS článku

—