A proof-of-concept that transfers distribution-level demographic parity metrics and a Wasserstein composite loss to outcome-oriented predictive process monitoring, with a tunable accuracy-fairness trade-off.
A Collection of Simulated Event Logs for Fairness Assessment in Process Mining
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
The analysis of fairness in process mining is a significant aspect of data-driven decision-making, yet the advancement in this field is constrained due to the scarcity of event data that incorporates fairness considerations. To bridge this gap, we present a collection of simulated event logs, spanning four critical domains, which encapsulate a variety of discrimination scenarios. By simulating these event logs with CPN Tools, we ensure data with known ground truth, thereby offering a robust foundation for fairness analysis. These logs are made freely available under the CC-BY-4.0 license and adhere to the XES standard, thereby assuring broad compatibility with various process mining tools. This initiative aims to empower researchers with the requisite resources to test and develop fairness techniques within process mining, ultimately contributing to the pursuit of equitable, data-driven decision-making processes.
citation-role summary
citation-polarity summary
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Achieving Group Fairness through Independence in Predictive Process Monitoring
A proof-of-concept that transfers distribution-level demographic parity metrics and a Wasserstein composite loss to outcome-oriented predictive process monitoring, with a tunable accuracy-fairness trade-off.