Repository of colleges and higher education institutions

Show document
A+ | A- | Help | SLO | ENG

Title:Model urejanja in popisovanja nestrukturiranih besedil z uporabo strojnega učenja : doktorska disertacija študijskega programa tretje bolonjske stopnje Arhivske znanosti
Authors:ID Milovanović, Miroslav (Author)
ID Novak, Miroslav (Mentor) More about this mentor... New window
Files:.pdf DR_Milovanovic_Miroslav_2024.pdf (5,79 KB)
MD5: 1AA46B1FD4BA81156CED8E2666E27D18
 
Language:Slovenian
Work type:Doctoral dissertation
Typology:2.08 - Doctoral Dissertation
Organization:UAMEU - Alma Mater Europaea University
Abstract:Namen: Namen doktorske disertacije je raziskati, ali je možna izdelava modela za urejanje in popisovanje nestrukturiranih besedil z uporabo strojnega učenja. Pri izdelavi modela je bila raziskava razdeljena na tri ključne segmente in povezana raziskovalna vprašanja, in sicer, ali je izdelava modela za samostojno klasifikacijo nestrukturiranih vsebin, samostojno prepoznavo imenskih entitet in samostojno izdelavo naslova popisne enote izvedljiva in uporabna. Metodologija: V raziskavi sta uporabljeni metoda analize vsebine in metoda eksperimenta. Raziskani so bili različni pristopi za izdelavo izvedbenega modela za urejanje in popisovanje nestrukturiranih besedil, ravno tako je bilo raziskana uporabnost izdelanega modela in učinkovitost izdelave popisne enote z uporabo izdelanega modela. Rezultati: Ugotovljeni rezultati raziskav kažejo, da je izdelava modela za urejanje in popisovanje nestrukturiranih besedil z uporabo strojnega učenja za vse tri segmente izvedljiva in uporabna, izdelani model pa predstavlja celoten formalni in aplikativni okvir za obdelavo nestrukturiranih besedil, ki se ga lahko neposredno uporabi za obdelavo nestrukturiranih podatkov. Izvirnost/uporabnost: Raziskava omogoča natančen vpogled v izdelavo modela za urejanje in popisovanje nestrukturiranih besedil ter izpostavlja prednosti in obliko uporabe izdelanega modela. Hkrati izdelani model in spremna dokumentiranost izdelave modela predstavljata podlago za uporabo modela v praksi in potencialno podlago za nadaljnje raziskave.
Keywords:urejanje in popisovanje, strojno učenje, imenske entitete, nestrukturirano besedilo, klasifikacija
Place of publishing:Maribor
Place of performance:Maribor
Publisher:M. Milovanović
Year of publishing:2024
Year of performance:2024
Number of pages:[10] f., 167 str.
PID:20.500.12556/ReVIS-11139 New window
COBISS.SI-ID:213858563 New window
UDC:930.25:004.85(043.3)
Publication date in ReVIS:09.12.2024
Views:201
Downloads:14
Metadata:XML DC-XML DC-RDF
:
Copy citation
  
Share:Bookmark and Share


Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Secondary language

Language:English
Abstract:Purpose: The purpose of the doctoral dissertation is to investigate whether it is possible to create a model for arrangement and archival description of unstructured texts using machine learning. When creating the model, the research was divided into three key segments and related research questions, namely whether the creation of a model for the standalone classification of unstructured texts, the standalone recognition of name entities and the standalone creation of the title of the unit of description is feasible and useful. Methodology: Methods content analysis and experiment are used in the research. Different approaches for creating an implementation model for arrangement and archival description of unstructured texts were investigated, as well as the usability of the developed model and the efficiency of creating a unit of description using the developed model. Results: The results of the research indicate that the creation of a model for arrangement and archival description of unstructured texts using machine learning is feasible and useful for all three segments, and the created model represents the entire formal and application framework for the processing of unstructured texts, which can be directly used for processing unstructured data. Originality/Usability: The research provides a detailed insight into the creation of a model for arrangement and archival description of unstructured texts and highlights the advantages and use cases of the created model. At the same time, the created model and the accompanying documentation of the model creation represent the basis for the use of the model in practice and a potential basis for further research.
Keywords:arrangement and description, machine learning, named entities, unstructured text, classification


Back