Ontology Alignment Techniques

View Sample PDF

Author(s): Marcos Martínez Romero (University of A Coruña, Spain), José Manuel Vázquez Naya (University of A Coruña, Spain), Javier Pereira Loureiro (University of A Coruña, Spain)and Norberto Ezquerra (Georgia Institute of Technology, USA)
Copyright: 2009
Pages: 6
Source title: Encyclopedia of Artificial Intelligence
Source Author(s)/Editor(s): Juan Ramón Rabuñal Dopico (University of A Coruña, Spain), Julian Dorado (University of A Coruña, Spain)and Alejandro Pazos (University of A Coruña, Spain)
DOI: 10.4018/978-1-59904-849-9.ch189

Keywords: Artificial Intelligence / Computer Science & IT / Information Science Reference

Purchase

View Ontology Alignment Techniques on the publisher's website for pricing and purchasing information.

Abstract

Sometimes the use of a single ontology is not sufficient to cover different vocabularies for the same domain, and it becomes necessary to use several ontologies in order to encompass the entire domain knowledge and its various representations. Disciplines where this occurs include medical science and biology, as well as many of its associated subfields such as genetics, epidemiology, etc. This may be due to a domain’s complexity, expansiveness, and/or different perspectives of the same domain on the part of different groups of users. In such cases, it is essential to find relationships that may exist between the elements of a specific domain’s different ontologies, a process known as ontology alignment. There are several methods for identifying the relationships or correspondences between elements associated with different ontologies, and collectively these methods are called ontology alignment techniques. Many of these techniques stem from other fields of study (e.g., matching techniques in discrete mathematics) while others have been specifically designed for this purpose. The key to successfully aligning ontologies is based on the appropriate selection and implementation of a set of those ontology alignment techniques best suited for a particular alignment problem. Ontology alignment is a complex, tedious, and time-consuming task, especially when working with ontologies of considerable size (containing, for instance, thousands of elements or more) and which have complex relationships between the elements (for example, a particular problem domain in medicine). Furthermore, the true potential of ontology alignment is realized when different information-exchange processes are integrated automatically, thereby providing the framework for reaching a suitable level of efficient interoperability between heterogeneous systems. The importance of automatically aligning ontologies has therefore been a topic of major interest in recent years, and recently there has been a surge in a variety of software tools dedicated to aligning ontologies in either a fully or partially automated fashion. Some of these tools —generally referred to as ontology alignment systems— have been the result of well known and respected research centers, including Stanford University and Hewlett Packard Laboratories, for instance. In Shvaiko & Euzenat, 2007, updated information is given regarding the currently available ontology alignment systems. Each ontology alignment system combines different alignment approaches along with its own techniques, such that correspondences between the different ontologies can be detected in the most complete, precise, and efficient manner. Since each system is based on its own approximation techniques, different systems yield different results, and therefore the quality of the results can vary among systems. Most of the alignment systems are oriented to solving problems of a general nature, since ontologies associated with a single domain share certain characteristics that set them apart from ontologies associated with other domains. Recently, some systems have emerged that are designed to align ontologies in a specific domain. An example is the SAMBO alignment system (Lambrix, 2006) in the biomedical domain. These and other domain-specific systems can produce excellent results (when used for the domains for which they were designed), but are generally not useful when applied to other domains.

The IRMA Community

Research IRM

Ontology Alignment Techniques

Purchase

Abstract

Related Content

IRMA Sponsors