Publication
LogPPL: A Tool for Probabilistic Process Mining
Martin Kuhn; Joscha Grüger; Christoph Matheja; Andrey Rivkin
In: Jochen De Weerdt; Giovanni Meroni; Han van der Aa; Karolin Winter (Hrsg.). ICPM Doctoral Consortium and Demo Track 2024. International Conference on Process Mining (ICPM-2024), 6th International Conference on Process Mining (ICPM 2024), located at ICPM 2024, October 14-18, Copenhagen, Denmark, Vol. 3783, CEUR Workshop Proceedings, 10/2024.
Abstract
This paper introduces LogPPL, a novel tool designed to bridge the gap between Data Petri Nets (DPNs)
and probabilistic programming, enabling the generation of event logs with statistical guarantees via
probabilistic program executions. LogPPL implements the transformation of DPNs into probabilistic
programs written in the WebPPL language, allowing to harness the power of simulation and inference
engines supplied for the WebPPL environment. Our tool simplifies the configuration of the DPN
simulation setup and allows for exporting both event logs in XES format as well as WebPPL files. LogPPL
capabilities are demonstrated through various scenarios, showcasing its potential to enhance process
mining tasks by offering rigorous statistical modeling and advanced simulation features. The tool’s
design, features, and performance are evaluated, highlighting its utility in both academic and industrial
settings.