models evolve at different levels of abstraction, from the requirements specification to development of the source code. The models underlying this process are related and their elements are usually mutually dependent. To preserve consistency and enable synchronization when models are altered due to evolution, the underlying model dependencies need to be established and maintained. As there is a potentially large number of such relations, this process should be automated for suitable scenarios. This paper introduces a tractable approach to automating identification and encoding of model dependencies that can be used for model synchronization. The approach first uses association rules to map types between models and different levels of ion. It then makes use of formal concept analysis (FCA) on attributes of extracted models to identify clusters of model elements.