Weighing the Pros and Cons: Process Discovery with Negative Examples
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Contemporary process discovery methods take as inputs only positive examples of process executions, and so they are one-class classification algorithms. However, we have found negative examples to also be available in industry, hence we propose to treat process discovery as a binary classification problem. This approach opens the door to many well-established methods and metrics from machine learning, in particular to improve the distinction between what should and should not be allowed by the output model. Concretely, we (1) present a formalisation of process discovery as a binary classification problem; (2) provide cases with negative examples from industry, including real-life logs; (3) propose the Rejection Miner binary classification procedure, applicable to any process notation that has a suitable syntactic composition operator; and (4) apply this miner to the real world logs obtained from our industry partner, showing increased output model quality in terms of accuracy and model size.
|Title of host publication||Business Process Management - 19th International Conference, BPM 2021, Proceedings|
|Editors||Artem Polyvyanyy, Moe Thandar Wynn, Amy Van Looy, Manfred Reichert|
|Publication status||Published - 2021|
|Event||19th International Conference on Business Process Management, BPM 2021 - Rome, Italy|
Duration: 6 Sep 2021 → 10 Sep 2021
|Conference||19th International Conference on Business Process Management, BPM 2021|
|Periode||06/09/2021 → 10/09/2021|
|Series||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
© 2021, Springer Nature Switzerland AG.
- Binary classification, Labelled event logs, Negative examples, Process mining