Automatic Representation of Geographical Data from a Semantic Point of View through a New Ontology and Classification Techniques

Providing Geographical Information Systems with mechanisms for processing geo‐data based on their semantics may help to solve problems like heterogeneity. This is because GIS could process geo‐data focusing on their meaning and not on their syntax and/or structure. An important aspect for achieving these objectives is the establishment of an automatic means of correspondence between geo‐data and their conceptualization in Higher Levels Ontologies (HLO). In this article, a new type of Ontology is proposed (Data‐Representation Ontology (DRO)). This Ontology describes the semantic embedded in geo‐data, which cannot be represented in current types of Ontologies. Across this Ontology, heterogeneous geographical data can be integrated in the semantic space contributing positively to the development of solutions for the problems of interoperability between heterogeneous systems. Likewise, we propose a new method for the automatic generation of the DRO and its interrelationships with HLO, based on pattern classification techniques. The experiments show that once the DRO is generated, the classifier can classify all data correctly. Thus, these data are semantically enriched. Moreover, this article shows how the topological relationships can enrich the semantics in the generated Ontology and increase the effectiveness of spatial analysis.