Uključivanje korpusa latinskih tekstova CroALa u bazu znanja latinskog jezika LiLa

master's thesis

• Title (english): Including the CroALa corpus of Latin texts into the LiLa knowledge base
• Author: Petar Soldo
• Mentor: Neven Jovanović (mentor)
• Committee member: Vladimir Rezar (predsjednik povjerenstva)
• Committee member: Teo Radić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Humanities and Social Sciences; (Department of Classical philology); Zagreb
• Defense date and country: 2021-05-17, Croatia
• Scientific / art field, discipline and subdiscipline: HUMANISTIC SCIENCES; Philology; Classical Philology

Abstract

Unatoč postojanju velikog broja raznih digitalnih alata i resursa za obradu i istraživanje latinskog jezika, njihova je iskoristivost i interoperabilnost skromna zbog značajnih razlika u formatima kojima se koriste. Upravo je to početna premisa projekta LiLa: Linking Latin, koji provodi milansko sveučilište Università Cattolica del Sacro Cuore i čiji je cilj povezati obilje jezičnih resursa i alata razvijenih za latinski i učiniti ih međusobno iskoristivima. Ovaj rad prikazuje spajanje korpusa tekstova hrvatskih latinista CroALa kao tekstualnog resursa u LiLa-u. Projekt je zamišljen u dvije faze: lematizacija CroALa-e i spajanje tekstova u LiLa-u. Budući da projekt nije dovršen, prikazuju se samo rezultati iz dijela prve faze projekta. Najprije se opisuje odabir reprezentativnog uzorka tekstova CroALa-e na kojima su testirani automatski lematizatori i gramatički tageri, a zatim slijedi opis dohvaćanja tekstova iz XML datoteka i njihova tokenizacija. Potom se opisuju alati za automatsku lematizaciju i tagiranje koji su korišteni te se prikazuju neobrađeni rezultati i (preliminarne) metode komparativne evaluacije uspješnosti pojedinih alata. U završnom dijelu rada najavljuje sam proces povezivanja CroALa-e s LiLa-om, a zadnje se poglavlje bavi primjenom ovog projekta u nastavi latinskog jezika.

Abstract (english)

Despite a large number of various digital tools and resources for processing and researching Latin, their usability and interoperability remains modest because they use substantially different formats. This is the main premise of the LiLa: Linking Latin project, developed by the Catholic University of the Sacred Heart in Milan. Its main goal is to link these various linguistic resources and tools developed for Latin and make them interoperable. This paper presents the inclusion of the CroALa corpus of texts by Croatian Latin writers into LiLa as a textual resource. The project is conceived in two phases: lemmatizing CroALa and linking the texts to LiLa. Because the project is still in progress, only the results of a part of the first phase are presented. Firstly, the process of choosing a representative sample of texts from CroALa is presented. The texts are selected for testing automatic lemmatizers and POS taggers. After that, the querying of the texts from their XML files and their tokenization are described. Then the tools used for lemmatization and tagging are addressed, together with the raw results and (preliminary) methods for performance evaluation. In the closing part of the paper, the actual linking of CroALa to LiLa is announced. The last chapter deals with the application of this project in teaching Latin in secondary schools.

• Keywords: CroALa; LiLa: Linking Latin; korpus; hrvatski latinisti; lematizacija latinskog; gramatičko tagiranje latinskog; jezikoslovni povezani otvoreni podatci
• Keywords (english): CroALa; LiLa: Linking Latin; corpus; Croatian Latinists; lemmatizing Latin; POS tagging Latin; Linked Linguistic Open Data (LLOD)
• Language: croatian
• URN:NBN: urn:nbn:hr:131:504508
• Study programme: Title: Latin Language and Literature (double major); specializations in: Teaching, Research; Course: Teaching; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra edukacije latinskog jezika i književnosti (magistar/magistra edukacije latinskog jezika i književnosti)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2021-05-17 19:32:22