ArchOnto, a CIDOC-CRM-Based Linked Data Model for the Portuguese Archives

Abstract

Archives are faced with great challenges due to the vast amounts of data they have to curate. New data models are required, and work is underway. The International Council on Archives is creating the RiC-CM (Records in Context), and there is a long line of work in museums with the CIDOC-CRM (CIDOC Conceptual Reference Model). Both models are based on ontologies to represent cultural heritage data and link them to other information. The Portuguese national archives hold a collection with over 3.5 million metadata records, described with the ISAD(G) standard. The archives are designing a new linked data model and a technological platform with applications for archive contributors, archivists, and the public. The current work extends CIDOC-CRM into ArchOnto, an ontology-based model for archives. The model defines the relevant archival entities and properties and will be used to migrate existing records. ArchOnto accommodates the existing ISAD(G) information and takes into account its implementation with current technologies. The model is evaluated with records from representative fonds.After the test on these samples, the model is ready to be populated with the semi-automatic transformation of the ISAD records. The evaluation of the model and the population strategies will proceed with experiments involving professional and lay users.

Representing and Validating Cultural Heritage Knowledge Graphs in CIDOC-CRM Ontology

“…The semantic migration process was applied according to both a predefined set of rules called "Mapping Description Rules" and both ISAD (G) and ISAAR recommendations. Other studies went with CIDOC-CRM after adding new classes and properties to its ontology to cover the source ontologies' entities, such as the level of description [12]. Another research [13] utilized CIDOC CRM and its extensions (CRMarchaeo, and CRMsci) to represent archaeological excavation activities and the archaeologists' observations.…”

“…The evaluation of the above mentioned mappings was limited to generating questions for researchers [11,15], evaluating a sample set of archival records [12], submitting a questionnaire to several categories of users [18], or checking inferred properties and SPARQL queries [14].…”

“…In CIDOC CRM versions prior to 6 [21], there were P82a and P82b properties to denote the beginning and end of a time span. Contrary to paper [12], where they used the "P78_is_identified_by" property for time spans. We have chosen the use of P82a and P82b properties, so these two properties were added to the used version of the CIDOC CRM ontology.…”

“…For instance, Figure 10 demonstrates how the birthdate and birthplace properties in COURAGE were represented by the E67 Birth entity in CIDOC-CRM. In contrast to [12], we did not create an instance of E41_Appellation as there was no other property we needed to assign to it. Figure 10 exemplifies the representation mechanism utilized for all entity properties.…”

An Evaluation of Graph Databases and Object-Graph Mappers in CIDOC CRM-Compliant Digital Archives

“…We present a comparison of OGM solutions to support graph databases and ontologies for the core data model of a transactional archival software system. Our data model is based primarily on ArchOnto-an ontology being developed as the data model for the digital archive [26]. However, since graph databases are flexible, other ontologies can be used in combination with ArchOnto, as long as their concepts are modeled using the OGM.…”