A statistical analysis approach to assist model transformation evolution

Model Driven Engineering (MDE) is essentially based in metamodel definition, model edition and the specification of model transformations (MT) among these. In many cases the development, evolution and adaptation of these transformations is still carried out without the support of proper methods and tools to reduce the effort and related costs to these activities. In this work, a novel model testing approach specifically designed to assist the engineer in model transformation evolution is presented. A statistical analysis of the actual behavior of the transformations is performed by means of the computation of well-known information extraction metrics. In order to assist the MT adaptation, a detailed interpretation of the possible results of those metrics is also presented. And finally, the results of applying this approach on a Model-Driven Reverse Engineering (MDRE) scenario defined in the context of the MIGRARIA project are discussed.