D 3 . 3 . 4 : Ontology distances for contextualisation

Distances between ontologies are useful for searching, matching or visualising ontologies. We study the various distances that can be defined across ontologies and provide them in a NeOn toolkit plug-in, OntoSim, which is a library of distances that can be used for recontextualising. The following partners have taken an active part in the work leading to the elaboration of this document, even if they might not have directly contributed to the writing of this document or its parts: Executive Summary Measuring distances between ontologies can be very important in various tasks for different purposes (finding an ontology to replace another, finding an ontology in which queries can be translated, finding people or businesses using similar ontologies). In the framework of this work package, we have defined recontextualisation as the connection of ontologies to context ontologies through alignment. So this amounts to finding on the web (through Watson or the Alignment server) ontologies aligned with the ontology to contextualise or ontologies which are close with respect to a particular distance. Hence, context-based matching can rely on ontology distances or similarities. Other uses of ontology measures within NeOn include finding related ontologies (like in Watson and Cupboard), visualising ontologies in metric space, measuring similarities between peers, or finding paths to answer queries. Hence there is no universal criterion for deciding if an ontology is close or far from another. There are also many possible ways to compute a similarity or a distance between ontologies and they may have different properties. The goal of this deliverable is twofold: (1) evaluating the validity of some of these measures, and (2) providing a library of such measures which can be used within NeOn. Concerning the first goal, we review a number of possible measures. We have separated these measures in two categories: measures using the content of the ontologies alone (the ontology space) and measures using existing alignment between ontologies (the alignment space). Measures working in the ontology space are rather classic: they compare the content of both ontologies in order to find how they are similar or dissimilar. This comparison can be based on terminology, structure or be semantically grounded; they can consider the ontology as a whole or only salient features such as key concepts. Moreover, there are different methods for aggregating a measure between ontology entities into a measure between ontologies. We consider the latter independently from the former. Measures working in …