Matching Relational Schemata to Semantic Web Ontologies
Polyxeni Katsiouli (University of Athens, Greece), Petros Papapanagiotou (University of Athens, Greece), Vassileios Tsetsos (University of Athens, Greece), Christos Anagnostopoulos (University of Athens, Greece) and Stathes Hadjiefthymiades (University of Athens, Greece)
Copyright: © 2009
The Semantic Web (SW; Berners-Lee, Hendler, & Lassila, 2001) is already in its implementation phase and an indication of this is the intense research and development activity in the area of SW tools and languages. SW is based on metadata, which describe the semantics of the Web content. SW envisages the enrichment of data with semantics in order to be machine understandable and enable knowledge reasoning. The core element for achieving such Web evolution is ontology: “Ontology is an explicit specification of a conceptualization” (Gruber).
Regarding the source schema, we assume a relational database (RDB) schema deployed on a typical relational database management system (RDBMS).
The conceptual schema (CS) of the target ontology (ONT) is expressed in a description logic (DL) language, due to the popularity of DLs in the SW community. DLs are knowledge representation languages (subsets of first-order logic) that express knowledge about concepts and conceptual hierarchies. An ontology expressed in DL language consists of concepts (classes) that can be described by various constructs and may have several restrictions (axioms). Concepts are categorized to primitive and defined concepts. Roles define binary relationships between concepts or between a concept and a datatype. Concepts and roles of an ontology can be both organized in hierarchical structures through the inclusion relation ⊑ (i.e., is-a, generalization, or subsumption). In OWL-DL, which is a sublanguage of the Web ontology language (OWL; Antoniou & van Harmelen, 2001), the term role is referred to as property. A property has a domain and range. When the domain and the range of a property are (primitive or defined) concepts, the property is called object-property. In case the range of a property is a literal (e.g., integer, string), the property is called datatype-property.
Key Terms in this Chapter
Semantic Web: The evolution of the current World Wide Web in a way that it is also machine understandable in addition to being human understandable.
Similarity Measure: An algorithmic method that calculates the degree of some aspect of similarity between two entities.
Schema Matching: The process of matching elements of a source data model (i.e., schema) with elements of a target data model
RONTO: A schema-matching methodology and tool that facilitates ontology population from relational data through a schema matching process.
Ontology Population: The process of creating instances for an ontology. This process usually involves linking various data sources to the elements of the ontology.
Ontology: A model (usually logic based) representing the main entities and their relationships within a domain of discourse.
Data Migration: The process of translating data from one format to another.