Improving Simplicity by Discovering Nested Groups in Declarative Models

Vlad Paul Cosma*, Axel Kjeld Fjelrad Christfort, Thomas T. Hildebrandt, Xixi Lu, Hajo A. Reijers, Tijs Slaats

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

2 Downloads (Pure)

Abstract

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.

Original languageEnglish
Title of host publicationAdvanced Information Systems Engineering - 36th International Conference, CAiSE 2024, Proceedings
EditorsGiancarlo Guizzardi, Flavia Santoro, Haralambos Mouratidis, Pnina Soffer
PublisherSpringer
Publication date2024
Pages440-455
ISBN (Print)9783031610561
DOIs
Publication statusPublished - 2024
Event36th International Conference on Advanced Information Systems Engineering, CAiSE 2024 - Limassol, Cyprus
Duration: 3 Jun 20247 Jun 2024

Conference

Conference36th International Conference on Advanced Information Systems Engineering, CAiSE 2024
Country/TerritoryCyprus
CityLimassol
Period03/06/202407/06/2024
SeriesLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14663 LNCS
ISSN0302-9743

Bibliographical note

Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

Keywords

  • Choices
  • DCR Graphs
  • Declarative
  • Nested Groups
  • Process Discovery
  • Simplicity

Cite this