Chapter

Business Process Management Workshops

Volume 66 of the series Lecture Notes in Business Information Processing pp 158-169

A Critical Evaluation Study of Model-Log Metrics in Process Discovery

  • Jochen De WeerdtAffiliated withDepartment of Decision Sciences and Information Management, Katholieke Universiteit Leuven
  • , Manu De BackerAffiliated withDepartment of Decision Sciences and Information Management, Katholieke Universiteit LeuvenDepartment of HABE, Hogeschool Gent, Universiteit GentDepartment of Management Information Systems, University of Antwerp
  • , Jan VanthienenAffiliated withDepartment of Decision Sciences and Information Management, Katholieke Universiteit Leuven
  • , Bart BaesensAffiliated withDepartment of Decision Sciences and Information Management, Katholieke Universiteit Leuven

* Final gross prices may vary according to local VAT.

Get Access

Abstract

The development of a well-defined evaluation framework for process discovery techniques is definitely one of the most important challenges within this subdomain of process mining. Any researcher in the field will acknowledge that such a framework is vital. With this paper, we aim to provide a tangible analysis of the currently available model-log evaluation metrics for mined control-flow models. Also, we will indicate strengths and weaknesses of the existing metrics and propose a number of opportunities for future research.

Keywords

process discovery evaluation metrics machine learning