6546 KB49 KB201920252025Sabadin, Ivančicaštevilka:1letnik:35Str. 45-69COBISSID_HOST:250077443ISSN:2670-4560URN:URN:NBN:SI:doc-VJVLGXUQenAlma Mater Europaea - European Center MariborInternational Institute for Archival ScienceAtlanti +archival descriptionarhivski opisKNIMEmachine learningRiC-O ontologijaRiC-O ontologysemantical enrichmentsemantična obogatitevstrojno učenjeTransforming archival description into semantically enriched form using machine learning|Purpose: The purpose of this paper is to determine if it is possible to transform the archival description in the relational database into an ontology with a ma-chine learning algorithm. Method/approach: The research will be based on the CRISP-ML(Q) method. The following steps will be carried out: Business and data understanding; Data preparation; Modelling and Evaluation.Results: After the transformation of the archival description, the Random Forest classification was used to predict the predicate in the semantic triplets. The re-sults obtained were: precision: 86.1% and accuracy: 96.5%.Conclusions / findings: Based on the results, we can conclude that the hypoth-esis was confirmed and that the machine learning algorithms are suitable for transforming the archival description in a structured form into an ontologyTEXTznanstveno časopisjejournalsSlovenian National E-content AggregatorNational and University Library of SloveniaMednarodni inštitut arhivskih znanosti Trst-Maribor