Tree Edit Distance Based Ontology Merging Evaluation Framework.

Merging structured knowledge has been widely investigated to build common resources in recent years. Indeed, many merging operators have been proposed and developed. However, the majority of them lack comparison and evaluation. Finding ontology sources for evaluation is not an effortless task. To this end, we propose a framework for evaluating the quality of ontology merging operators. The primary strategy starts with an original ontology as a gold standard to create noisy ontologies as datasets and use them to evaluate the merging operators. We generate the noisy ontologies using some perturbations of the tree structure of the original ontology based on tree edit operations. Then, we use tree edit distance to measure the existing merging operators with these noisy sources. We provide the details to assess the merging operators' efficiency in the computation time and their ability to cover (or be close to) the original ontology.