Uncertain process data with probabilistic knowl- edge: Problem characterization and challenges

Izack Cohen, Avigdor Gal

Research output: Contribution to journalConference articlepeer-review

2 Scopus citations


Motivated by the abundance of uncertain event data from mul- tiple sources including physical devices and sensors, this paper presents the task of relating a stochastic process observation to a process model that can be rendered from a dataset. In contrast to previous research that suggested to transform a stochastically known event log into a less informative uncertain log with upper and lower bounds on activity frequencies, we consider the challenge of accommodating the probabilistic knowledge into conformance checking techniques. Based on a taxonomy that captures the spectrum of conformance checking cases under stochastic process observations, we present three types of challenging cases. The first includes conformance checking of a stochastically known log with respect to a given process model. The second case extends the first to classify a stochastically known log into one of several process models. The third case extends the two previous ones into settings in which process models are only stochastically known. The suggested problem captures the increasingly growing number of applications in which sensors provide probabilistic process information.

Original languageEnglish
Pages (from-to)51-56
Number of pages6
JournalCEUR Workshop Proceedings
StatePublished - 2021
Event2021 International Workshop on BPM Problems to Solve Before We Die, PROBLEMS 2021 - Rome, Italy
Duration: 6 Sep 202110 Sep 2021

Bibliographical note

Publisher Copyright:
© 2021 CEUR-WS. All rights reserved.


  • Conformance checking
  • Process classification
  • Sensors
  • Stochastically known traces


Dive into the research topics of 'Uncertain process data with probabilistic knowl- edge: Problem characterization and challenges'. Together they form a unique fingerprint.

Cite this