Automatic root cause identification using most probable alignments

Marie Koorneef, Andreas Solti, Henrik Leopold*, Hajo A. Reijers

*Corresponding author for this work

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

In many organizational contexts, it is important that behavior conforms to the intended behavior as specified by process models. Non-conforming behavior can be detected by aligning process actions in the event log to the process model. A probable alignment indicates the most likely root cause for non-conforming behavior. Unfortunately, available techniques do not always return the most probable alignment and, therefore, also not the most probable root cause. Recognizing this limitation, this paper introduces a method for computing the most probable alignment. The core idea of our approach is to use the history of an event log to assign probabilities to the occurrences of activities and the transitions between them. A theoretical evaluation demonstrates that our approach improves upon existing work.

Original languageEnglish
Title of host publicationBusiness Process Management Workshops - BPM 2017 International Workshops, Revised Papers
PublisherSpringer/Verlag
Pages204-215
Number of pages12
ISBN (Electronic)9783319740300
ISBN (Print)9783319740294
DOIs
Publication statusPublished - 2018
Event15th International Conference on Business Process Management, BPM 2017 - Barcelona, Spain
Duration: 10 Sept 201715 Sept 2017

Publication series

NameLecture Notes in Business Information Processing
Volume308
ISSN (Print)1865-1348

Conference

Conference15th International Conference on Business Process Management, BPM 2017
Country/TerritorySpain
CityBarcelona
Period10/09/1715/09/17

Keywords

  • Conformance checking
  • Most probable alignments
  • Root cause analysis

Fingerprint

Dive into the research topics of 'Automatic root cause identification using most probable alignments'. Together they form a unique fingerprint.

Cite this