Contribución a la alineación de ontologías utilizando lógica difusa

Abstract

With the increasing amount of information available on the Internet today, it is becoming critical to create mechanisms to facilitate the organization and to enable the exchange of information and knowledge between applications. The Semantic Web is intended to solve one of the fundamental limitations of the current Web: the lack of ability for representations to express meanings. This task can be simplified greatly by adding semantic information and context to current forms of knowledge representation used in the Web, so that computers can process, interpret and connect the information in the WWW. Ontologies have become a crucial component in the Semantic Web, allowing the design of exhaustive and rigorous conceptual schemas to facilitate communication and information exchange between different systems and institutions. However, the heterogeneity in knowledge representation in ontologies hampers the interaction between the applications that make use of this knowledge. Therefore, to share information between applications using heterogeneous vocabularies, they must be able to translate data from one ontological framework to another. The process of finding correspondences between different ontologies is called ontology alignment. This Ph.D. thesis proposes an ontology alignment method using fuzzy logic techniques to combine a set of similarity measures between different ontologies entities. The proposed similarity measures are based on two fundamental elements of ontologies: the terminology and structure. Regarding terminology we propose a linguistic similarity measure using a set of lexical relations between the names of the entities combined with a semantic similarity measure that takes into account the context information in ontology entities. In terms of structure, we propose similarity measures which use both relational structure and the internal structure of concepts within ontologies.