Detecting Logical Couplings at Model Level by Analyzing EMF Model repository

In Model-driven engineering (MDE), models constitute the basis source of information to specify and automatically generate an executable system. The evolution history analysis of software system models helps us to discover interesting logical couplings between its artifacts; which is not possible with the analysis of a single version of the model. The detection of such couplings earlier in the software design improves the model (software) in a way to increase its modularity. In this paper, we propose an approach for the analysis of a version repository for EMF models (CDO Model Repository) in order to detect logical couplings between the software model artifacts, using the Apriori algorithm. Such logical couplings help us to understand the model evolution, to predict future changes on the model, to identify potential faults, and consequently to produce automatically a coherent and modular software source code.