Ontology Mapping

Declares that two classes or properties from different ontologies have the same meaning and extension. This is the strongest mapping relationship.

• Use Case: Merging duplicate concepts, such as mapping foaf:Person to schema:Person.
• Formal Logic: Implies mutual subsumption (C ≡ D).
• Impact: Enables query federation; a query for one concept automatically retrieves instances of the other.

Defines a hierarchical is-a relationship, stating that all instances of a source class are also instances of a more general target class.

• Directional: Mapping is explicit about which concept is more specific.
• Example: Mapping a proprietary Internal:Manager class to the standard foaf:Person class.
• Consequence: Enables inheritance of properties and supports hierarchical reasoning.

Establishes correspondences between properties (relationships) and constrains their applicability.

• Core Task: Mapping ex:employeeName to foaf:name.
• Domain/Range Alignment: Ensuring the classes that the properties link are compatible. Mapping fails if ex:employeeName has a domain of ex:Machines while foaf:name has a domain of foaf:Person.
• Critical for: Accurate data transformation and preserving relational semantics.

Asserts that two individual instances (nodes) from different graphs refer to the same real-world entity.

• Foundation for: Entity resolution and deduplication across datasets.
• Example: Linking dbpedia:Berlin to wikidata:Q64.
• Challenges: Requires sophisticated similarity metrics and validation to avoid erroneous merges, which can propagate errors through reasoning.

Maps a concept in one ontology to a logical expression built from concepts in another, using Boolean operators.

• Intersection (owl:intersectionOf): Maps Source:StudentEmployee to Target:Student and Target:Employee.
• Union (owl:unionOf): Maps Source:Asset to Target:Vehicle or Target:Building.
• Requires: A reasoning engine to interpret the logical expression and correctly classify instances.

Defines that two mapped concepts cannot have any instances in common. This prevents incorrect inference and maintains data integrity.

• Critical for: Error detection in mapping. If Source:Organization is mapped to Target:Person, but the target ontology states Person and Organization are disjoint, the mapping is inconsistent.
• Use Case: Enforcing business rules, such as ensuring a Contract is not also a PhysicalProduct.

The process of establishing semantic correspondences between the attributes, tables, or classes of two or more heterogeneous data schemas. It is a prerequisite for ontology mapping, focusing on structural similarities.

• Key Technique: Often uses linguistic, structural, and instance-based matchers.
• Example: Aligning a Customer table in a CRM with a Client entity in an ERP system.
• Outcome: Produces a set of attribute-to-attribute mappings, which inform the deeper concept-to-concept mappings of ontology mapping.

The declarative process of defining rules to transform raw, heterogeneous data (CSV, JSON, XML, databases) into RDF triples conforming to a target ontology. It operationalizes ontology mappings.

• Standard: RDF Mapping Language (RML) is a W3C-standardized extension of R2RML for non-relational sources.
• Function: Maps source fields to ontology classes (rdf:type), properties, and literals.
• Tooling: Implemented by engines like SDM-RDFizer and Morph-KGC to automate knowledge graph population.

The process of determining whether records from different sources refer to the same real-world entity, a critical step before applying ontology mappings to instance data.

• Core Challenge: Disambiguating "John Smith, NYC" from "J. Smith, New York".
• Techniques: Uses fuzzy matching, rule-based logic, and machine learning for similarity scoring.
• Integration: Resolved entities are then mapped to a unified canonical node in the knowledge graph, enabling coherent linking.

The broader process of standardizing data from disparate sources by resolving syntactic, structural, and semantic differences. Ontology mapping is its semantic component.

• Scope: Encompasses format conversion, unit standardization, code list alignment, and semantic mapping.
• Goal: Creates a consistent, comparable dataset where revenue_USD and income_GBP are transformed and understood as equivalent financial measures.
• Business Impact: Essential for integrated reporting, analytics, and compliance.

The process of converting data that has multiple possible representations into a single, standard, authoritative form. It enforces the results of ontology mapping at the data instance level.

• Example: Converting "New York," "NY," "N.Y." and "New York City" to a canonical URI <http://example.org/geo/NewYorkCity>.
• Role: Works in tandem with mapping to ensure all data adhering to a mapped concept uses identical identifiers and value formats.
• Benefit: Eliminates ambiguity and ensures deterministic query results.

A data integration methodology that applies semantic technologies—ontologies, RDF, and mapping languages—within an ETL pipeline to produce a semantically consistent knowledge graph.

• Transform Stage: The T in ETL is driven by ontology mappings and RML rules.
• Output: Produces linked RDF data instead of or in addition to traditional tabular formats.
• Key Differentiator: The transformation is meaning-aware, preserving relationships and context lost in traditional ETL.