A method towards process similarity measurement according to constrained traces

The wide application of Business Process Management has led to the existence of a tremendous amount of process models. Similarity measurement of process models is usually considered a versatile solution to fulfill their effective utilization. Although many researches haves worked on different notions of model similarity, most of them are not precise enough, as there is always a gap between the structural and behavioral similarity measurements and the impacts of some important activities are reduced because of the undifferentiated treatment on traces. To address the problem, in this paper, we introduce a proper metric to quantify process similarity based on constrained trace alignment. It is grounded in the constrained longest common subsequences which can retain the behavior and structure information of the original model and guarantee the priority of some significant parts of a process. The experiment result shows that this method can significantly improve the similarity value and make it more appropriate for practical application.