Replay using Recomposition: Decomposed Alignment Conformance


In the area of process mining, efficient alignment-based conformance checking is a hot topic. Existing approaches for conformance checking are typically monolithic and compute exact fitness values. One limitation with monolithic approaches is that it may take a significant amount of computation time in large processes. Alternatively, decomposition approaches run much faster but do not always compute an exact fitness value. This paper presents a tool Replay using Recomposition which returns the exact fitness value and the resulting alignments using the decomposition approach in an iterative manner. Other than computing the exact fitness value, users can configure the balance between result accuracy and computation time to get a fitness interval within set constraints, e.g., “Give me the best fitness estimation you can find within 5 minutes.”


Following you can see some screenshots of the application:

Parameters Setting View


Resulting Alignments View



To illustrate the use of Recomposing Conformance a screencast can be seen here. This video includes a walkthrough of the different components of the tool and an example of decomposed alignment-based conformance checking analysis.


Example datasets can be found in


  • Verbeek, H.M.W., Lee, J., Munoz-Gama, J., van der Aalst, W.M.P., Sepúlveda (2017). Replay using Recomposition: Alignment-BasedConformance Checking in the Large (under review)
  • Lee, J., Verbeek, H.M.W., Munoz-Gama, J., van der Aalst, W.M.P., Sepúlveda (2017). Recomposing Conformance: Closing the Circle on Decomposed Alignment-Based Conformance Checking in Process Mining (under review)


This work is partially supported by CONICYT-PCHA/Doctorado Nacional/2017-21170612, by the Vicerrectoría de Investigacion de la Pontificia Universidad Católica de Chile/Concurso Estadías y Pasantías Breves 2016, and by the Departamento de Ciencias de la Computación UC/Fond-DCC-2017-0001.