Element level semantic matching using WordNet

We think of Match as an operator which takes two graph-like structures and produces a mapping between semantically related nodes. The matching process is essentially divided into two steps: element level and structure level. Element level matchers consider only labels of nodes, while structure level matchers start from this information to consider the full graph. In this paper we present and evaluate, on the large scale real world dataset, twelve new element level semantic matchers. The matchers exploit WordNet as a background knowledge source, and return semantic relations (e.g., equivalence, more general) between concepts rather than similarity coefficients between labels in the [0, 1] range. The twelve element level semantic matchers are evaluated against seven state of the art matching systems. The results of the matchers are found comparable with the results of the matching systems.