An Instance-based Learning Approach for Ontology Matching

This paper proposes an instance-based learning approach for the ontology matching problem. This approach is applicable to scenarios where instances of the ontologies to be matched are exchanged between sources. An initial population of instances is used as a training set of a non-supervised algorithm that constructs mappings between properties of classes from the ontologies. To demonstrate the e ectiveness of our approach, we carried out a set of experiments in which we varied the size of the training set for two di erent ontologies. Our results showed that our approach achieves low false positive rates for su cient large datasets, although it is totally dependent on the heterogeneity of the domains of properties.