TAGER: Transition-Labeled Graph Edit Distance Similarity Measure on Process Models

摘要

Although several approaches have been proposed to compute the similarity between process models, they have various limitations. We propose an approach named TAGER (Transition-lAbeled Graph Edit distance similarity MeasuRe) to compute the similarity based on the edit distance between coverability graphs. As the coverability graph represents the behavior of a Petri net well, TAGER, based on it, has a high precise computation. Besides, the T-labeled graphs (an isomorphic graph of the coverability graph) of models are independent, so TAGER can be used as the index for searching process models in a repository. We evaluate TAGER from efficiency and quality perspectives. The results show that TAGER meets all the seven criteria and the distance metric requirement that a good similarity algorithm should have. TAGER also balances the efficiency and precision well.