New quality metrics for evaluating process models

Zan Huang, Akhil Kumar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations


In the context of business process intelligence, along with the need to extract a process model from a log, there is also the need to measure the quality of the extracted process model. Hence, process model quality notions and metrics are required. We present a systematic approach for developing quality metrics for block structured process models, which offer less expressive power than Petri-nets but have easier semantics. The metrics are based on tagging an initial block structured process model with self-loop and optional markings in order to explain all the instances in the given log. Then we transform the marked model to an equivalent maximal model by rewriting the self-loop and optional markings for consistency, and determine a badness score for it, which determines quality. Our approach is compared with related work, and a plan for testing and validation on noise-free and noisy data is discussed.

Original languageEnglish (US)
Title of host publicationBusiness Process Management Workshops - BPM 2008 International Workshops - Revised Papers
PublisherSpringer Verlag
Number of pages7
ISBN (Print)9783642003271
StatePublished - Jan 1 2009
Event6th International Conference on Business Process Management - Workshops, BPM 2008 - Milano, Italy
Duration: Sep 1 2008Sep 4 2008

Publication series

NameLecture Notes in Business Information Processing
Volume17 LNBIP
ISSN (Print)1865-1348


Other6th International Conference on Business Process Management - Workshops, BPM 2008

All Science Journal Classification (ASJC) codes

  • Management Information Systems
  • Control and Systems Engineering
  • Business and International Management
  • Information Systems
  • Modeling and Simulation
  • Information Systems and Management


Dive into the research topics of 'New quality metrics for evaluating process models'. Together they form a unique fingerprint.

Cite this