TY - GEN
T1 - Improving Simplicity by Discovering Nested Groups in Declarative Models
AU - Cosma, Vlad Paul
AU - Christfort, Axel Kjeld Fjelrad
AU - Hildebrandt, Thomas T.
AU - Lu, Xixi
AU - Reijers, Hajo A.
AU - Slaats, Tijs
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024/6/3
Y1 - 2024/6/3
N2 - Discovering simple, understandable and yet accurate process models is a well-known issue for models mined from real-life event logs. In this paper, we consider algorithms for automatically computing nested groups of activities in declarative process languages, concretely Dynamic Condition Response (DCR) Graphs, to reduce complexity while preserving accuracy. The DCR Graphs notation is, on the one hand, supported by the very accurate DisCoveR process mining algorithm, and on the other hand, by mature design and execution tools used in industrial processes and enterprise information management systems. We evaluate our approach by applying the DisCoveR miner to a large benchmark of real-life and synthetic event logs, measuring the size, density, separability, and constraint variability of mined models with and without grouping of activities. In earlier work, these measures have been shown to have a significant effect on the intrinsic cognitive load for users of declarative models, in particular DCR Graphs. We also evaluate the effect of prioritizing in particular the grouping of activities that model mutual exclusive choices. Our evaluation confirms that grouping of activities in general lowers the complexity on 3 of the 4 measures, while prioritizing choices in some cases makes the improvement slightly smaller.
AB - Discovering simple, understandable and yet accurate process models is a well-known issue for models mined from real-life event logs. In this paper, we consider algorithms for automatically computing nested groups of activities in declarative process languages, concretely Dynamic Condition Response (DCR) Graphs, to reduce complexity while preserving accuracy. The DCR Graphs notation is, on the one hand, supported by the very accurate DisCoveR process mining algorithm, and on the other hand, by mature design and execution tools used in industrial processes and enterprise information management systems. We evaluate our approach by applying the DisCoveR miner to a large benchmark of real-life and synthetic event logs, measuring the size, density, separability, and constraint variability of mined models with and without grouping of activities. In earlier work, these measures have been shown to have a significant effect on the intrinsic cognitive load for users of declarative models, in particular DCR Graphs. We also evaluate the effect of prioritizing in particular the grouping of activities that model mutual exclusive choices. Our evaluation confirms that grouping of activities in general lowers the complexity on 3 of the 4 measures, while prioritizing choices in some cases makes the improvement slightly smaller.
KW - Choices
KW - DCR Graphs
KW - Declarative
KW - Nested Groups
KW - Process Discovery
KW - Simplicity
UR - http://www.scopus.com/inward/record.url?scp=85196753558&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-61057-8_26
DO - 10.1007/978-3-031-61057-8_26
M3 - Conference contribution
AN - SCOPUS:85196753558
SN - 978-3-031-61056-1
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 440
EP - 455
BT - Advanced Information Systems Engineering - 36th International Conference, CAiSE 2024, Proceedings
A2 - Guizzardi, Giancarlo
A2 - Santoro, Flavia
A2 - Mouratidis, Haralambos
A2 - Soffer, Pnina
PB - Springer
T2 - 36th International Conference on Advanced Information Systems Engineering, CAiSE 2024
Y2 - 3 June 2024 through 7 June 2024
ER -