Automatic classification to matching patterns for process model matching evaluation
Elena Kuss, Heiner Stuckenschmidt · MADOC (University of Mannheim) · 2017
Business process model matching is concerned with the detection of similarities in business process models.To support the progress of process model matching techniques, efficient evaluation strategies are required.State-of-the-art evaluation techniques provide a grading of the evaluated matching techniques.However, they only offer limited information about strength and weaknesses of the individual matching technique.To efficiently evaluate matching systems, it is required to automatically analyze the attributes of the matcher output.In this paper, we propose an evaluation by automatic classification of the alignments to matching patterns.On the one hand, to understand strength and weaknesses of a matching technique.On the other hand, to identify potential for further improvement.Consequently, optimal matching scenarios of a specific matcher can be derived.This further enables tuning of a matcher to specific applications.