BPI Challenges

Since process mining is a data driven research field, real-life datasets have always been a cornerstone of the work in this field. To convince companies to share their datasets publicly and to convince researchers to actively use these datasets, the BPI challenge was first organized in 2011. Co-located with the BPI Workshop at the time, the aim of the BPI Challenge was to allow both researchers and practitioners to showcase their work.

Over the years, the BPI Challenge datasets have become important benchmarks in the community. Some datasets, such as the financial logs from 2012 and 2017, have been studied in detail by many researchers and students, while others are used to highlight the usefulness of specific techniques, such as concept drift or the discovery of multi-instance patterns.

This page provides information on the past challenge and, as soon as data becomes available, details on the current challenge which, since 2019, is co-located with the task-force's flagship conference: ICPM.