Similarity features, and their role in concept alignment learning

Finding mappings between compatible ontologies is an important and difficult open problem. Instance-based methods for solving this problem have the advantage of focussing on the most active parts of the ontologies and reflect the semantics of the ontologies as they are used in the real world. We evaluate how the feature representation of the instances is representative of the corresponding concepts, investigate how this corresponds with the domain characteristics of the data and which role it plays in the task of instance-based ontology mapping. We use two different competitive classifiers and a standard feature selection to identify important features, and study the effect of those different classifiers in the concept alignment context.