InstanceMatching.org

Instance Matching is the activity of evaluating the degree of similarity among different descriptions of real objects across heterogeneous data sources. Quite often, the goal of Instance Matching is to determine whether two object descriptions are referred to the same real object in a given domain, under the assumption that the higher is the similarity between two descriptions, the higher is the probability that the two descriptions are actually referred to the same object. In computer science, Instance Matching is also the activity of defining methods, techniques and (semi-)automated tools for performing the similarity evaluation task. Moreover, we use the term Instance Matching to focus the attention on the matching task dealing with Semantic Web data sources (e.g., OWL aboxes, RDF documents, SCORM data).