A Spectrum of Entropy-Based Precision and Recall Measurements Between Partially Matching Designed and Observed Processes

Abstract

Modern software systems are often built using service-oriented principles. Atomic components, be that web-or micro services, allow constructing flexible and loosely coupled systems. In such systems, services are building blocks orchestrated by business processes the system supports. Due to the complexity and heterogeneity of industrial software systems, implemented processes may deviate from those initially designed. In this paper, we propose a spectrum of conformance measurements. The spectrum results from a generalization of the recently introduced entropy-based approaches for measuring precision and recall between observed process executions and designed process models. The new generalized measures of precision and recall inherit the desired for this class of measures properties and provide analysts with flexible control over the sensitivity for identifying commonalities and discrepancies in the compared processes and performance of the techniques. The reported evaluation based on our implementation of the measures over real-world event logs and automatically discovered models confirms the feasibility of using the approach in industrial settings.