SKOS (Simple Knowledge Organization System)

SKOS revolves around the skos:Concept class, which represents a unit of thought or idea. Unlike OWL, which focuses on formal class definitions and logical axioms, SKOS describes semi-formal conceptual resources using labels and documentation properties.

• Each concept is a distinct URI resource, enabling unambiguous reference across distributed systems.
• Concepts are linked through semantic relationship properties rather than rigid logical constraints.
• This model is ideal for representing existing thesauri, taxonomies, and subject heading systems in a machine-readable format.

SKOS provides three distinct properties for attaching human-readable labels to concepts, supporting multilingual vocabularies and user interface display.

• skos:prefLabel: The single preferred lexical label for a concept in a given language, used for display and indexing.
• skos:altLabel: Alternative labels including synonyms, abbreviations, and quasi-synonyms that facilitate search recall.
• skos:hiddenLabel: Labels accessible for string matching but not visible to end users, useful for handling common misspellings or deprecated terms.

Concepts are connected through standardized relationship properties that define the structure of the knowledge organization system.

• skos:broader and skos:narrower: Assert hierarchical parent-child relationships, enabling tree-like navigation and query expansion.
• skos:related: Links conceptually associated concepts that are not hierarchically related, such as a discipline and its object of study.
• skos:broaderTransitive and skos:narrowerTransitive: Super-properties that enable inference of indirect hierarchical ancestry without requiring explicit assertion of every transitive link.

SKOS includes a rich set of annotation properties for documenting concept scope, usage, and editorial history, bridging the gap between informal knowledge organization and formal ontology.

• skos:scopeNote: Clarifies the intended meaning and boundaries of a concept within the vocabulary.
• skos:definition: Provides a formal textual definition of the concept's meaning.
• skos:example: Supplies illustrative instances of the concept's application.
• skos:historyNote: Records significant changes in the meaning or status of the concept over time.
• skos:editorialNote: Captures internal administrative information for vocabulary maintainers.

SKOS provides dedicated properties for establishing correspondences between concepts in different knowledge organization systems, enabling semantic interoperability without requiring full ontology alignment.

• skos:exactMatch: Asserts that two concepts have identical meaning and can be used interchangeably.
• skos:closeMatch: Indicates sufficiently similar concepts that can be substituted in many applications.
• skos:broadMatch and skos:narrowMatch: Express hierarchical mappings across distinct schemes.
• skos:relatedMatch: Links associated concepts across different vocabularies.
• These properties are critical for Linked Data integration and cross-walking between thesauri like AGROVOC and EUROVOC.

SKOS organizes concepts into aggregations using two distinct constructs that serve different structural purposes.

• skos:ConceptScheme: Represents a complete thesaurus, classification, or vocabulary as a single entity, providing a container for its constituent concepts and metadata like title and publisher.
• skos:Collection: Groups concepts that share a common characteristic without implying hierarchical relationships, useful for node labels in systematic displays or thematic groupings.
• skos:OrderedCollection: A subtype of Collection where the sequence of members is meaningful, supporting guided navigation paths.
• skos:inScheme: Explicitly links each concept to the scheme it belongs to, ensuring provenance and context.

The computational process of determining correspondences between heterogeneous ontologies to achieve semantic interoperability, also known as ontology matching. When two organizations model the same domain using different SKOS concept schemes, alignment techniques identify equivalences (e.g., skos:exactMatch) between their concepts. Key approaches include:

• Lexical matching using string similarity metrics on skos:prefLabel
• Structural matching comparing hierarchical skos:broader relationships
• Extensional matching comparing instance sets linked via skos:related

A high-level, domain-independent framework defining abstract, philosophical categories such as time, space, and object. Upper ontologies like BFO (Basic Formal Ontology) partition reality into continuants and occurrents, providing a common integration hub. SKOS concept schemes often map their top-level concepts to upper ontology classes to facilitate broad semantic interoperability between domain-specific knowledge bases without requiring direct pairwise alignment.

A core OWL property that asserts two named individuals refer to the exact same real-world entity. In SKOS-based Linked Data, owl:sameAs forms the critical identity link for interlinking distributed concept schemes. For example, asserting dbpedia:Artificial_Intelligence owl:sameAs wikidata:Q11660 merges the identity of the concept across datasets. Caution: misuse of owl:sameAs for near-synonyms can introduce logical inconsistencies; prefer skos:closeMatch or skos:exactMatch for conceptual linking.

A technique that learns low-dimensional vector representations for entities across different knowledge graphs in a unified space, where geometric proximity indicates semantic equivalence. When applied to SKOS-structured taxonomies, embedding alignment can discover latent cross-walks between concept schemes that lack explicit skos:mappingRelation links. Models like MTransE and IPTransE jointly optimize for structure preservation and alignment accuracy across graphs.

A query answering regime that evaluates SPARQL queries not just against explicitly asserted triples but against the full logical closure of the RDF graph derived from inference rules. For SKOS data, entailment regimes can materialize implicit hierarchical relationships—for instance, inferring transitive skos:broader paths across multiple levels of a taxonomy. Supported regimes include:

• RDFS entailment for basic subclass and subproperty reasoning
• SKOS-specific entailment for skos:broaderTransitive closure