Modern infrastructures for information and communication technologies are aimed at providing enhanced services by integrating the knowledge spread on the web through an ontological representation of information. However, ontology usefulness in managing different knowledge sources is limited by the so-called semantic heterogeneity problem arising when several interacting software components use different ontologies for representing the same information. In order to bridge this gap and, consequently, enable a full interoperability across the software components, it is necessary to bring the corresponding ontologies into a mutual agreement by identifying a set of semantic relationships among their entities. This result is achieved by means of a so-called ontology alignment process that, for each pair of entities belonging to the ontologies under alignment, computes their semantic closeness through an optimized aggregation of different similarity measures. Unfortunately, this similarity aggregation is a hard optimization process, above all, when no information is known about ontology features. The aim of this paper is to define an ontology alignment process based on a memetic algorithm able to efficiently aggregate similarity measures without using a priori knowledge about ontologies under alignment. As shown by a statistical multiple comparison procedure, our approach yields high performance in terms of alignment quality with respect to top-performers of well-known Ontology Alignment Evaluation Initiative campaigns.
Enhancing ontology alignment through a memetic aggregation of similarity measures
LOIA, Vincenzo;VITIELLO, AUTILIA
2013-01-01
Abstract
Modern infrastructures for information and communication technologies are aimed at providing enhanced services by integrating the knowledge spread on the web through an ontological representation of information. However, ontology usefulness in managing different knowledge sources is limited by the so-called semantic heterogeneity problem arising when several interacting software components use different ontologies for representing the same information. In order to bridge this gap and, consequently, enable a full interoperability across the software components, it is necessary to bring the corresponding ontologies into a mutual agreement by identifying a set of semantic relationships among their entities. This result is achieved by means of a so-called ontology alignment process that, for each pair of entities belonging to the ontologies under alignment, computes their semantic closeness through an optimized aggregation of different similarity measures. Unfortunately, this similarity aggregation is a hard optimization process, above all, when no information is known about ontology features. The aim of this paper is to define an ontology alignment process based on a memetic algorithm able to efficiently aggregate similarity measures without using a priori knowledge about ontologies under alignment. As shown by a statistical multiple comparison procedure, our approach yields high performance in terms of alignment quality with respect to top-performers of well-known Ontology Alignment Evaluation Initiative campaigns.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.