doctoral thesis
Modelling authority metadata in Resource Descrition Framework

Predrag Perožić (2016)
University of Zadar
Metadata
TitleModeliranje autoriziranih metapodataka u Općem okviru za opisivanje resursa
AuthorPredrag Perožić
Mentor(s)Mirna Willer
Abstract
Doktorski rad „Modeliranje autoriziranih metapodataka u Općem okviru za opisivanje resursa“ jeste istraživanje najprikladnije mogućnosti da se bibliografski autorizirani metapodaci pohranjeni u UNIMARC/A formatu predstave kao semantički povezani i strojno razumljivi podaci pomoću podatkovnog modela kojeg nazivamo Opći okvir za opisivanje resursa (Resource Description Framework, RDF). Rezultat doktorskog rada je konkretan proizvod pod nazivom Ontologija autoriziranih podataka - OAP, koja se zasniva na apstraktnoj sintaksi SKOSXL/RDF podatkovnog modela i na semantici koja je modelirana prema specifičnim logičkim podatkovnim elementima UNIMARC/A formata koji opisuju bibliografske entitete. Skup elemenata OAP ontologije sastoji se od sto i dvije (102) klase i sto četrdeset i sedam (147) svojstava te je modularno proširen Ontologijom administrativnih podataka - ADM. Skup elemenata ADM ontologije sastoji se od jedne (1) klase i šesnaest (16) svojstava koja su modelirana prema specifičnim logičkim podatkovnim elementima UNIMARC/A formata koji opisuju UNIMARC/A zapis. Elementi OAP i ADM ontoloških rječnika, ili skraćeno jednostavno OAP, postižu istu razinu granulacije autoriziranih metapodataka kao i znatno veći broj UNIMARC/A označitelja sadržaja (kombinacija polje+indikatori+potpolje). Budući da je OAP model jednostavniji alat od UNIMARC/A formata ali omogućava saopćavanje istih informacija možemo reći da je postignuta svojevrsna optimizacija. U metodološkom smislu OAP ontologija kombinira induktivni (bottom-up) i deduktivni (top-down) pristup i pokušava postići ravnotežu između: a) jasnoće i konzistentnosti u definiranju bibliografskih entiteta kao OAP klasa i njihovih semantičkih odnosa kao OAP svojstava, i b) preciznosti i količine informacija koje opisuju te entitete u skupu autoriziranih podataka. Doktorski rad osim teorijskog obrazloženja sadrži i cjelovitu specifikaciju OAP ontologije u ljudski i strojno čitljivom obliku u OWL/XML notaciji čime su zadovoljeni teorijski i tehnički preduvjeti za njenu implementaciju.
KeywordsOntologija autoriziranih podataka (OAP) UNIMARC Authorities Format authority data modelling RDF SKOS SKOSXL
Parallel title (English)Modelling authority metadata in Resource Descrition Framework
Committee MembersMarijana Tomić (committee chairperson)
Mirna Willer (committee member)
Boris Bosančić (committee member)
GranterUniversity of Zadar
PlaceZadar
StateCroatia
Scientific field, discipline, subdisciplineSOCIAL SCIENCES
Information and Communication Sciences
UDK02
GENERALLY
Librarianship
Study programme typeuniversity
Study levelpostgraduate
Study programmeKnowledge Society and Information Transfer
Academic title abbreviationdr. sc.
Genredoctoral thesis
Language Croatian
Defense date2016-04-15
Parallel abstract (English)
The research addressed in this thesis, "Modelling authorized metadata in the Resource Description Framework," investigated the most suitable options, using the Resource Description Framework (RDF) data model, for representing authorized bibliographic metadata stored in UNIMARC/A format as semantically related and machine understandable data. The outcome of the research is a concrete product, Ontologija autoriziranih podataka (the Authority Data Ontology or OAP). OAP is based on the abstract syntax of the SKOSXL/RDF data model, and semantics that were modeled according to UNIMARC/A format-specific logical data elements describing bibliographic entities. The OAP element set consists of one hundred and two (102) classes and one hundred and forty seven (147) properties that are modularly extended by applying the Ontologija administrativnih podataka (Ontology of administrative data or ADM). The ADM element set consists of one (1) class and sixteen (16) properties that are modeled according to the UNIMARC/A format-specific logical data elements describing UNIMARC/A record. The OAP + ADM ontological vocabulary, which is abbreviated simply as OAP, achieves the same granularity of authorized metadata as well as a significantly higher number of UNIMARC/A content designators (a combination of field+indicators+subfield). Since the OAP vocabulary is a simpler tool than the UNIMARC/A format but transmits the same information it can be said that this approach achieves a kind of optimization. In terms of methodology, the OAP ontology combines both bottom-up and top-down approaches in trying to strike a balance between: a) clarity and consistency in defining bibliographic entities as OAP classes and their semantic relations as OAP properties, and b) the accuracy and amount of information that describes these entities in the authorized data set. In addition to its theoretical component, the thesis contains the full specification of the OAP ontology in human and machine-readable form in OWL/XML notation and thus meets the theoretical and technical conditions for its implementation.
Parallel keywords (Croatian)Ontologija autoriziranih podataka (OAP) UNIMARC format za pregledne zapise modeliranje autoriziranih podataka RDF SKOS SKOSXL
Extent259 str., 30 cm
Versionaccepted version
Resource typetext
Access conditionOpen access
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
Noteaccepted version
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:162:437195
CommitterMarta Unković