Collaborative Metadata Definition using Controlled Vocabularies, and Ontologies

Data's role in a variety of technical and research areas is undeniably growing. This can be seen, for example, the increased investments development data-intensive analytical methods such as artificial intelligence (Zhang 2022), well rising rate data generation which expected to continue into near future (Rydning Shirer 2021). Academic one areas, where lifeblood generating hypotheses, creating new knowledge, reporting results. Unlike proprietary industry data, academic often subjected stricter requirements regarding transparency, accessibility. part due public funding many institutions receive. One way fulfil these by observing FAIR (Findability, Accessibility, Interoperability, Reusability) principles scientific (Wilkinson et al. 2016). These introduce benefits, reproducibility, more transparent use funding, environmental sustainability. A implementing practice with help Digital Objects (FDOs) (European Commission: Directorate-General Research Innovation 2018). FDO consists an accompanying Persistent Identifier (PID), rich metadata describes context data. Additionally, format contained should widely used, ideally open. Our presentation focused on third FDO's components mentioned previously – metadata. It outlines concept framework enables collaborative definition fields used annotate FDO-encapsulated given domain research. The first component presented controlled vocabulary related needs annotated. collective that denotes list terms , their definitions relations between them. In this contribution, correspond annotation process. Formally, type vocabularies thesaurus (National Information Standards Organization 2010). Thesauri consist not only elements previously, but also allow inclusion synonyms every defined term. eliminates ambiguity occur when using similar definitions. thesauri specify simple hierarchical vocabulary, provide explicit structure set fields. most important feature our framework, however, developed fashion experts field. Specifically, people are able propose term edits, cast votes appropriateness have already been proposed. Despite advantages, limit lacking capability relating each other semantically fashion. motivated second namely ontologies. An ontology “a specification conceptualization” (Gruber 1995). More precisely, it represents entities domain, various After has within transformed contains additional extend beyond contain domain-specific information about For relation denote value field must take. Furthermore, ontologies link metadata, individual FDOs themselves. contribute Reusability aspect Data Objects. generated group linked existing ontology. Afterwards, reused easily researchers from same field, because will specified subject area. cross-domain combined increase reusability boundaries. described above being implemented form multiple software tools framework. one, editor written Python-based web application called VocPopuli, entry point who want develop or lab. software, whose version tested internally, definition, editing terms. annotates term, entire PROV Model (PROV-DM) (Moreau Missier 2013) - schema describe provenance object. Finally, assigns PID itself. worth noting themselves seen through prism FDOs: they (the terms) annotated (e.g., terms' authors) provided PID. solution facilitate transformation VocPopuli handle two distinct cases from-scratch conversion ontologies, augmentation thesaurus. As case tool Python programming language. solutions finally semi-overlapping groups users materials science. On hand, input, edit, discuss area interest, thus create vocabularies. administrators oversee creation, processes semi-automatic complete, creation experimental procedures, and/or integration richer augment published work (Garabedian 2022) thereby test resources. domains, science, tribology, metalworking. templates procedures above.