A Method for Recommending Ontology Alignment Strategies

Abstract

Abstract. In different areas ontologies have been developed and many of these ontologies contain overlapping information. Often we would therefore want to be able to use multiple ontologies. To obtain good results, we need to find the relationships between terms in the different ontologies, i.e. we need to align them. Currently, there already exist a number of different alignment strategies. However, it is usually difficult for a user that needs to align two ontologies to decide which of the different available strategies are the most suitable. In this paper we propose a method that provides recommendations on alignment strategies for a given alignment problem. The method is based on the evaluation of the different available alignment strategies on several small selected pieces from the ontologies, and uses the evaluation results to provide recommendations. In the paper we give the basic steps of the method, and then illustrate and discuss the method in the setting of an alignment problem with two well-known biomedical ontologies. We also experiment with different implementations of the steps in the method.

Putting Ontology Alignment in Context: Usage Scenarios, Deployment and Evaluation in a Library Case

“…While in the translation scenario, Silas, which detects links based on extensional information of concepts, performs as well as Falcon does. This is in line with the current trend in alignment research that investigates ways to perform case-specific selection of alignment strategies [19,20]. This also gives further reasons to keep up application-specific evaluation efforts in OAEI-like campaigns.…”

A Session-Based Approach for Aligning Large Ontologies

“…It is also clear that the approaches using validation decisions, become better the more suggestions are validated. For the approaches using segment pairs, the choice of the segment pairs influences the recommendation results (which is different from the conclusions of experiments in [19]). …”

“…We implemented several recommendation strategies. The first approach (an extension of [19]) requires the user or an oracle to validate all pairs in small segments of the ontologies. To generate these segments we first use a string-based approach to detect concepts with similar names.…”

Evaluating the Stability and Credibility of Ontology Matching Methods

“…In this way, these measures fail to assess a matching method based on the two important criteria, which makes users hard to select the most suitable method or system. Some methods [5,13,14] for selecting the best matching strategy take several parameters into account, yet these parameters do not include scores for evaluating stability and credibility.…”