Developing a core ontology for taxonomic data.

Kennedy, Jessie; Gales, Robert; Hyam, Roger; Kukla, Robert; Wieczorek, John; Hagedorn, Gregor; D�ring, Markus; Vieglais, Dave

Abstract

Over recent years several sub-groups within the Taxonomic Databases Working Group (TDWG) have developed models and exchange standards to facilitate data sharing within the taxonomic community. These include ABCD, SDD, DwC, TCS and Spatial Data Standards. Of these, ABCD, SDD and TCS have been ratified as TDWG standards (see http://www.tdwg.org/standards). Although each group focused on different aspects of taxonomic data and its representation, the resulting standards duplicated the modeling of many aspects of taxonomy. For example, Biological Collections refer to specimens, taxonomic names or concepts, institutions, people, publications; Descriptions refer to people, specimens, publications, taxonomic names and concepts etc; and Taxonomic Concepts refer to specimens, taxonomic names, publications, people, descriptions etc. The resulting overlap across the existing standards has limited or no common terminology or model. A Core Ontology is proposed as a reference for all taxonomic domain models to facilitate more effective data sharing within the community. Representatives from ABCD, DwC, GBIF, SDD and TCS analysed the existing data models and exchange standards and propose a draft Core Ontology for taxonomy. The result is expressed in terms of a Base Ontology, Core Ontology and Domain Ontology from which applications can be developed. The Base Ontology comprises classes representing general, non-taxonomic specific concepts which are seen as base classes from which the Core Ontology classes are derived. The Core Ontology comprises classes that correspond to the most common and therefore important concepts within the TDWG community. These classes are seen as the basis of a community vocabulary. Where possible, such classes were given textual definitions based on Oxford English dictionary to aid in the general understanding of what was intended by the class. The properties of Core Ontology classes are limited in type other classes in the Core Ontology, leaving further elaboration to the domain classes. A Domain Ontology is developed from the classes in Core Ontology. The Domain Ontology comprises sub-ontologies which have a correspondence to a single class in the Core Ontology to encourage reusability of the Domain Ontology classes. This, for example, would prevent a Specimen ontology defining Taxonomic Name. The Domain Ontology classes capture additional semantics to the core classes. This allows the community to share classes and further to compose application schemas to represent their perspective of the taxonomic domain such as museum curation, ecological surveys and nomenclature management. Relationships introduced between classes not explicit in the Core Ontology may indicate misuse or misunderstanding of the semantics of the Core Ontology. Application schemas will develop classes/data structures using the classes and the properties in the Domain Ontology. The presentation will report on the design and development of the Core Ontology and how the Core was extended to a trial Domain Ontology. The trial Domain Ontology had an associated data repository that enabled the testing of the Core Ontology.

Citation

Kennedy, J., Gales, R., Hyam, R., Kukla, R., Wieczorek, J., Hagedorn, G., Döring, M., & Vieglais, D. (2006, October). Developing a core ontology for taxonomic data. Presented at Taxonomy Data Working Group (TDWG) 2006

• Conference Name: Taxonomy Data Working Group (TDWG) 2006
• Start Date: Oct 15, 2006
• End Date: Oct 22, 2006
• Publication Date: 2006
• Deposit Date: Oct 26, 2009
• Publicly Available Date: Oct 26, 2009
• Peer Reviewed: Peer Reviewed
• Book Title: Proceedings of TDWG (2006), St Louis, MI
• Keywords: Biology; Taxonomy; Data exchange standards; Overlapping terms; Duplicated terms; Proposals; Base Ontology; Core Ontology; Domain Ontology; Methodology; Application; Case study
• Public URL: http://researchrepository.napier.ac.uk/id/eprint/3031
• Contract Date: Oct 26, 2009