Metamodel matching techniques evaluation and benchmarking

During the last decades, the matching operation has been thoroughly studied in database systems and ontology development. Matching is critical in many application domains such as semantic web, schema and data integration, data translation. In the Model Driven Engineering (MDE) context, matching techniques between two metamodels are the centerpiece of a semi-automatic transformation process. Metamodel matching allows discovering mappings between two metamodels and then generating transformation rules. In this paper, we propose an approach for metamodel matching techniques evaluation and benchmarking. This approach firstly provides a benchmark for metamodel matching and, secondly prepares to semi-automate the transformation process according to the Model Driven Architecture (MDA). A benchmark is useful to compare the strengths and weaknesses of different metamodel matching techniques. Experimental results comparing the effectiveness of the application of various implementations of four recent techniques (Similarity Flooding, ModeICVS, SAMT4MDE andAML) on real-world metamodels are given.