Knowledge engineering: principles and methods
Data & Knowledge Engineering - Special jubilee issue: DKE 25
Intelligent Information Integration For The Semantic Web (Lecture Notes in Computer Science)
Intelligent Information Integration For The Semantic Web (Lecture Notes in Computer Science)
The Dissimilarity Representation for Pattern Recognition: Foundations And Applications (Machine Perception and Artificial Intelligence)
Comparing categories among geographic ontologies
Computers & Geosciences
Spatial relations for semantic similarity measurement
ER'05 Proceedings of the 24th international conference on Perspectives in Conceptual Modeling
| Hi-index | 0.00 |
Providing Geographical Information Systems (GIS) with the mechanisms for processing geographical data based on their semantic abstraction is a task that at present is carried out in a number of research given their scope of applications. Tackling this issue may help to solve many problems of geographical data like its heterogeneity, since the SIG could process geographical data focusing on their meaning and not on their syntax and/or structure, thus reducing the Man-Machine semantic gap. An important aspect for achieving these objectives is the establishment of an automatic way of correspondence between geographical data and their conceptualization in a Domain Ontology. In this work, we propose a new type of Ontology, a Data-Representation Ontology. We also propose a new method for the automatic generation of the Data-Representation Ontology from geographical data and his interrelationships with the Domain Ontology. For this we use pattern classification techniques and a dissimilarity measure. The experiments showed that once the Data-Representation Ontology was generated, the classifier using dissimilarities could correctly classify all the data.