A unified approach for measuring precision and generalization based on anti-alignments

B.F. Van Dongen, J. Carmona, T. Chatain

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

13 Citations (Scopus)
2 Downloads (Pure)

Abstract

The holy grail in process mining is an algorithm that, given an event log, produces fitting, precise, properly generalizing and simple process models. While there is consensus on the existence of solid metrics for fitness and simplicity, current metrics for precision and generalization have important flaws, which hamper their applicability in a general setting. In this paper, a novel approach to measure precision and generalization is presented, which relies on the notion of antialignments. An anti-alignment describes highly deviating model traces with respect to observed behavior. We propose metrics for precision and generalization that resemble the leave-one-out cross-validation techniques, where individual traces of the log are removed and the computed anti-alignment assess the model’s capability to describe precisely or generalize the observed behavior. The metrics have been implemented in ProM and tested on several examples.

Original languageEnglish
Title of host publicationBusiness Process Management
Subtitle of host publication14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings
EditorsB. La Rosa, P. Loos, O. Pastor
Place of PublicationDordrecht
PublisherSpringer
Pages39-56
Number of pages18
ISBN (Electronic)978-3-319-45348-4
ISBN (Print)978-3-319-45347-7
DOIs
Publication statusPublished - 2016

Publication series

NameLecture notes in computer science
Volume9850
ISSN (Print)0302-9743

Fingerprint

Alignment
Metric
Trace
Process Mining
Cross-validation
Process Model
Fitness
Simplicity
Defects
Generalise
Generalization
Model

Cite this

Van Dongen, B. F., Carmona, J., & Chatain, T. (2016). A unified approach for measuring precision and generalization based on anti-alignments. In B. La Rosa, P. Loos, & O. Pastor (Eds.), Business Process Management: 14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings (pp. 39-56). (Lecture notes in computer science; Vol. 9850). Dordrecht: Springer. https://doi.org/10.1007/978-3-319-45348-4_3
Van Dongen, B.F. ; Carmona, J. ; Chatain, T. / A unified approach for measuring precision and generalization based on anti-alignments. Business Process Management: 14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings. editor / B. La Rosa ; P. Loos ; O. Pastor. Dordrecht : Springer, 2016. pp. 39-56 (Lecture notes in computer science).
@inproceedings{572e4efaad2947f196e805f063a00584,
title = "A unified approach for measuring precision and generalization based on anti-alignments",
abstract = "The holy grail in process mining is an algorithm that, given an event log, produces fitting, precise, properly generalizing and simple process models. While there is consensus on the existence of solid metrics for fitness and simplicity, current metrics for precision and generalization have important flaws, which hamper their applicability in a general setting. In this paper, a novel approach to measure precision and generalization is presented, which relies on the notion of antialignments. An anti-alignment describes highly deviating model traces with respect to observed behavior. We propose metrics for precision and generalization that resemble the leave-one-out cross-validation techniques, where individual traces of the log are removed and the computed anti-alignment assess the model’s capability to describe precisely or generalize the observed behavior. The metrics have been implemented in ProM and tested on several examples.",
author = "{Van Dongen}, B.F. and J. Carmona and T. Chatain",
year = "2016",
doi = "10.1007/978-3-319-45348-4_3",
language = "English",
isbn = "978-3-319-45347-7",
series = "Lecture notes in computer science",
publisher = "Springer",
pages = "39--56",
editor = "{La Rosa}, B. and P. Loos and O. Pastor",
booktitle = "Business Process Management",
address = "Germany",

}

Van Dongen, BF, Carmona, J & Chatain, T 2016, A unified approach for measuring precision and generalization based on anti-alignments. in B La Rosa, P Loos & O Pastor (eds), Business Process Management: 14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings. Lecture notes in computer science, vol. 9850, Springer, Dordrecht, pp. 39-56. https://doi.org/10.1007/978-3-319-45348-4_3

A unified approach for measuring precision and generalization based on anti-alignments. / Van Dongen, B.F.; Carmona, J.; Chatain, T.

Business Process Management: 14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings. ed. / B. La Rosa; P. Loos; O. Pastor. Dordrecht : Springer, 2016. p. 39-56 (Lecture notes in computer science; Vol. 9850).

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

TY - GEN

T1 - A unified approach for measuring precision and generalization based on anti-alignments

AU - Van Dongen, B.F.

AU - Carmona, J.

AU - Chatain, T.

PY - 2016

Y1 - 2016

N2 - The holy grail in process mining is an algorithm that, given an event log, produces fitting, precise, properly generalizing and simple process models. While there is consensus on the existence of solid metrics for fitness and simplicity, current metrics for precision and generalization have important flaws, which hamper their applicability in a general setting. In this paper, a novel approach to measure precision and generalization is presented, which relies on the notion of antialignments. An anti-alignment describes highly deviating model traces with respect to observed behavior. We propose metrics for precision and generalization that resemble the leave-one-out cross-validation techniques, where individual traces of the log are removed and the computed anti-alignment assess the model’s capability to describe precisely or generalize the observed behavior. The metrics have been implemented in ProM and tested on several examples.

AB - The holy grail in process mining is an algorithm that, given an event log, produces fitting, precise, properly generalizing and simple process models. While there is consensus on the existence of solid metrics for fitness and simplicity, current metrics for precision and generalization have important flaws, which hamper their applicability in a general setting. In this paper, a novel approach to measure precision and generalization is presented, which relies on the notion of antialignments. An anti-alignment describes highly deviating model traces with respect to observed behavior. We propose metrics for precision and generalization that resemble the leave-one-out cross-validation techniques, where individual traces of the log are removed and the computed anti-alignment assess the model’s capability to describe precisely or generalize the observed behavior. The metrics have been implemented in ProM and tested on several examples.

UR - http://www.scopus.com/inward/record.url?scp=84989350535&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-45348-4_3

DO - 10.1007/978-3-319-45348-4_3

M3 - Conference contribution

AN - SCOPUS:84989350535

SN - 978-3-319-45347-7

T3 - Lecture notes in computer science

SP - 39

EP - 56

BT - Business Process Management

A2 - La Rosa, B.

A2 - Loos, P.

A2 - Pastor, O.

PB - Springer

CY - Dordrecht

ER -

Van Dongen BF, Carmona J, Chatain T. A unified approach for measuring precision and generalization based on anti-alignments. In La Rosa B, Loos P, Pastor O, editors, Business Process Management: 14th International Conference, BPM 2016, Rio de Janeiro, Brazil, September 18-22, 2016. Proceedings. Dordrecht: Springer. 2016. p. 39-56. (Lecture notes in computer science). https://doi.org/10.1007/978-3-319-45348-4_3