Abstract
Geospatial ontologies have a key role to play in the development of the geospatial-Semantic Web, with regard to facilitating the search for geographical information and resources. They normally hold large volumes of geographic information and undergo a continuous process of revision and update. Limitations of the OWL ontology representation language for supporting geospatial domains are discussed and an integrated rule and ontology language is recognized as needed to support the representation and reasoning requirements in this domain. A survey of the current approaches to integrating ontologies and rules is presented and a new framework is proposed that is based on and extends Description Logic Programs. A hybrid representational approach is adopted where the logical component of the framework is used to represent geographical concepts and spatial rules and an external computational geometry processor is used for storing and manipulating the associated geometric data. A sample application is used to demonstrate the proposed language and engine and how they address the identified challenges.
TopManaging Geospatial Ontologies
In this section we consider a typical geospatial ontology model, as shown in Figure 1. The model is based on OGC guidelines for simple geographic features, see (OGC Technical Committee, 1999; Vretanos, 2005), and other models commonly used in existing geospatial ontology development (Fu et al., 2005, Smith & Frew, 1995). The terminology of the geo-ontology is relatively plain with regards to the number and type of constructs used. This reflects typical geographic ontology developments which, beyond the complex representation of geometry, are relatively sparse (parsimonious (Jones et al., 2001)) and fit to purpose. In this section, issues related to the representation and management of such geospatial ontologies are discussed and OWL’s ability to handle them is evaluated.
Figure 1. An example geospatial ontology
Key Terms in this Chapter
Ontology: Those things that exist are those things that have a formal representation within the context of a machine. Knowledge commits to an ontology if it adheres to the structure, vocabulary and semantics intrinsic to a particular ontology i.e. it conforms to the ontology definition. A formal ontology in computer science is a logical theory that represents a conceptualization of real world concepts.
Interleaved Mixed Mode Reasoning: Allowing knowledge querying using backward rule sets during the course of forward inferencing. Facts in the premise of forward rules can be found directly from explicit facts or implicitly through additional inferences.
Description Logics: Successors to Semantic Networks and Frame Based Languages which can represent both asserted and structural knowledge. Modern description logics stem from KL-ONE which formalized the ideas of Semantic Networks and Frames. A description logic can describe the world in terms of properties or constraints that specific individuals have to satisfy.
Geospatial Rules: Rules or logical rules of inference represent dynamic, relational knowledge as opposed to static structural knowledge. A rule is an inference of the form PREMISE implies CONCLUSION, mimicking human cognitive reasoning processes. If the PREMISE condition holds, then the CONCLUSION condition is deducible. Geospatial rules are those that represent and reason with geographical and or spatial knowledge.
Integration of Rules and Ontologies: How to integrate description logic ontologies with classical rule based logic programs, while preserving semantics and maintaining decidability and tractability. Approaches are either based on the union or the intersection of the two languages.
Geospatial Ontologies: A specilisation of ontology that represent only knowledge from the geographic and spatial domain.
Qualitative Spatial Reasoning: Representation of continuous properties of the world by discrete symbols, and then reasoning over such symbols without recourse to more expensive (computationall) quantitative knowledge. Qualitative knowledge and reasoning better mimics human spatial reasoning processes.