Business process model matching: An approach based on semantics and structure

Process model similarity identification has recently gained an increasing interest in the business process management domain. It can be used in business process model refactoring, retrieval, etc. Existing approaches use the structural aspect of process models hence producing imprecise matching results. We present in this paper a semi-automatic approach to match business process model elements based on both their semantics and their structure. Our approach first applies a set of semantic rules to detect semantic relationships among activity labels, e.g. like equivalence, subsumption and partial-correspondence. Secondly, it uses the structure of matched process elements to identify simple and complex correspondences between process model fragments. As illustrated through a set of examples, our approach has a promising recall and precision rates.