Article Preview
TopIntroduction
Over the last several years, diverse ontologies have been created to enable interoperability across multiple and diverse information systems, especially with the advent of the Semantic Web. Ontologies are at the kernel of Semantic Web by allowing the explicitness of the semantic purpose for structuring different fields of interest. They play a key role in annotating web pages or services by modeling the concepts, attributes and relationships used to annotate resource content.
Nowadays the use of ontologies in many contexts such as science, e-commerce, e-health, and industry has been increased (Stuckenschmidt, 2009). With the rapid development of ontology applications, domain ontologies became very large in scale, which generates several difficulties in ontology constructing, matching, reusing, maintenance, reasoning, etc.
In many application contexts, several ontologies covering the same or related fields are developed independently of each other by different communities, which raises the issue of being able to exchange, integrate and transform data. At this stage, the problem of interoperability arises, allowing heterogeneous systems to communicate and cooperate, and to this end, semantic links must be established between entities belonging to two different ontologies, and the transition to the web is a real challenge that requires researchers to make efforts to optimize content management, which can be constantly enriched and developed. To this end, it is necessary to improve the quality of the organization, structuring, research, identification, access, use, integration, and automated processing of this content. In the past several years, a panoply of systems has been developed to automatically generate, match, and integrate ontologies in a process called ontology alignment. In each of these cases, it is important to know the relationships between the terms in the different ontologies. It has been realized that this is a major issue and much research has recently been done on ontology alignment, i.e. finding mappings between terms in different ontologies.
Due to their sizes and monolithic nature, large ontologies lead to a new challenge to the state-of-the-art ontology alignment technology. Ontology alignment is an important way of establishing interoperability between applications that use different but related ontologies (Hu et al.,2008).
In the past several years, an enormous effort has been invested in aligning and harmonizing ontologies including rule-based and statistical matchers. Existing matchers rely on entity features such as names, synonyms, as well as relationships to other entities (Shvaiko and Euzenat, 2013; Otero-Cerdeira et al., 2015). However, manually aligning ontologies is a labor-intensive process and often semi-automatic approaches are sought.
All alignment techniques are required to scale up to handle large ontologies (Jimenez-Ruiz et al., 2012; Diallo, 2014). However, this is not always obvious because the creation of multiple ontologies, sometimes for the same domain, leads to heterogeneity between the knowledge expressed within each of them that must be resolved: it is the problem of interoperability.
When ontologies cover overlapping topics, the overlap can be represented using ontology alignments. These alignments need to be continuously adapted to changing ontologies. Especially for large ontologies this is a costly task often consisting of manual work (Jurisch and Igler, 2018).
The research field of ontology alignment is very active with its yearly workshop as well as a yearly event, the Ontology Alignment Evaluation Initiative (OAEI, e.g. (Euzenat, 2011)), that focuses on evaluating systems that automatically generate mapping suggestions. OAEI systems are typically able to cope with small and medium size ontologies, but fail to complete large tasks in a given time frame and/or with the available resources (e.g., memory). Large ontologies still pose serious challenges to state-of-the-art ontology alignment systems (Jimenez-Ruiz et al., 2018).