Can deep neural networks learn process model structure? An assessment framework and analysis
(2022)
PROCESS MINING WORKSHOPS, ICPM 2021.
In Lecture Notes in Business Information Processing
433.
p.127-139
- Author
- Jari Peeperkorn, Seppe vanden Broucke (UGent) and Jochen De Weerdt
- Organization
- Abstract
- Predictive process monitoring concerns itself with the prediction of ongoing cases in (business) processes. Prediction tasks typically focus on remaining time, outcome, next event or full case suffix prediction. Various methods using machine and deep learning have been proposed for these tasks in recent years. Especially recurrent neural networks (RNNs) such as long short-term memory nets (LSTMs) have gained in popularity. However, no research focuses on whether such neural network-based models can truly learn the structure of underlying process models. For instance, can such neural networks effectively learn parallel behaviour or loops? Therefore, in this work, we propose an evaluation scheme complemented with new fitness, precision, and generalisation metrics, specifically tailored towards measuring the capacity of deep learning models to learn process model structure. We apply this framework to several process models with simple control-flow behaviour, on the task of next-event prediction. Our results show that, even for such simplistic models, careful tuning of overfitting countermeasures is required to allow these models to learn process model structure.
- Keywords
- Predictive Process Monitoring, Next Event Prediction, Recurrent Neural Network, Generalisation
Downloads
-
Process Mining Workshops.pdf
- full text (Published version)
- |
- open access
- |
- |
- 538.03 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-8751963
- MLA
- Peeperkorn, Jari, et al. “Can Deep Neural Networks Learn Process Model Structure? An Assessment Framework and Analysis.” PROCESS MINING WORKSHOPS, ICPM 2021, vol. 433, Springer, 2022, pp. 127–39, doi:10.1007/978-3-030-98581-3_10.
- APA
- Peeperkorn, J., vanden Broucke, S., & De Weerdt, J. (2022). Can deep neural networks learn process model structure? An assessment framework and analysis. PROCESS MINING WORKSHOPS, ICPM 2021, 433, 127–139. https://doi.org/10.1007/978-3-030-98581-3_10
- Chicago author-date
- Peeperkorn, Jari, Seppe vanden Broucke, and Jochen De Weerdt. 2022. “Can Deep Neural Networks Learn Process Model Structure? An Assessment Framework and Analysis.” In PROCESS MINING WORKSHOPS, ICPM 2021, 433:127–39. Cham: Springer. https://doi.org/10.1007/978-3-030-98581-3_10.
- Chicago author-date (all authors)
- Peeperkorn, Jari, Seppe vanden Broucke, and Jochen De Weerdt. 2022. “Can Deep Neural Networks Learn Process Model Structure? An Assessment Framework and Analysis.” In PROCESS MINING WORKSHOPS, ICPM 2021, 433:127–139. Cham: Springer. doi:10.1007/978-3-030-98581-3_10.
- Vancouver
- 1.Peeperkorn J, vanden Broucke S, De Weerdt J. Can deep neural networks learn process model structure? An assessment framework and analysis. In: PROCESS MINING WORKSHOPS, ICPM 2021. Cham: Springer; 2022. p. 127–39.
- IEEE
- [1]J. Peeperkorn, S. vanden Broucke, and J. De Weerdt, “Can deep neural networks learn process model structure? An assessment framework and analysis,” in PROCESS MINING WORKSHOPS, ICPM 2021, 2022, vol. 433, pp. 127–139.
@inproceedings{8751963,
abstract = {{Predictive process monitoring concerns itself with the prediction of ongoing cases in (business) processes. Prediction tasks typically focus on remaining time, outcome, next event or full case suffix prediction. Various methods using machine and deep learning have been proposed for these tasks in recent years. Especially recurrent neural networks (RNNs) such as long short-term memory nets (LSTMs) have gained in popularity. However, no research focuses on whether such neural network-based models can truly learn the structure of underlying process models. For instance, can such neural networks effectively learn parallel behaviour or loops? Therefore, in this work, we propose an evaluation scheme complemented with new fitness, precision, and generalisation metrics, specifically tailored towards measuring the capacity of deep learning models to learn process model structure. We apply this framework to several process models with simple control-flow behaviour, on the task of next-event prediction. Our results show that, even for such simplistic models, careful tuning of overfitting countermeasures is required to allow these models to learn process model structure.}},
author = {{Peeperkorn, Jari and vanden Broucke, Seppe and De Weerdt, Jochen}},
booktitle = {{PROCESS MINING WORKSHOPS, ICPM 2021}},
isbn = {{9783030985806}},
issn = {{1865-1348}},
keywords = {{Predictive Process Monitoring,Next Event Prediction,Recurrent Neural Network,Generalisation}},
language = {{eng}},
pages = {{127--139}},
publisher = {{Springer}},
title = {{Can deep neural networks learn process model structure? An assessment framework and analysis}},
url = {{http://doi.org/10.1007/978-3-030-98581-3_10}},
volume = {{433}},
year = {{2022}},
}
- Altmetric
- View in Altmetric
- Web of Science
- Times cited: