pith. sign in

arxiv: 2212.11047 · v4 · submitted 2022-12-21 · 💻 cs.DB

Discovering Process Models With Long-Term Dependencies While Providing Guarantees and Filtering Infrequent Behavior Patterns

Lisa Luise Mannel , Wil M. P. van der Aalst This is my paper

Pith reviewed 2026-05-24 10:36 UTC · model grok-4.3

classification 💻 cs.DB
keywords process discoveryPetri netseST-Minerfitness guaranteeplace fitness metricevent logsinfrequent behavior
0
0 comments X

The pith

Adaptations to the eST-Miner select place subsets so the resulting Petri net guarantees a minimum fitness on the event log while preserving high precision.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper modifies the eST-Miner so that places evaluated separately are chosen in a way that their combination into one net still replays at least a chosen fraction of the input traces. It targets the problem that independent place checks can produce deadlocks or low overall fitness once transitions are shared. A revised fitness measure for single places is added to stop infrequent labels from being excluded while still ignoring rare trace patterns. Experiments on real and synthetic logs compare how different selection rules and metrics affect the final models.

Core claim

By adapting place selection inside the eST-Miner, a subset of places can be chosen so the assembled Petri net meets a user-specified minimum fitness on the event log; a refined place fitness metric is defined and tested that reduces blocking of infrequent activity labels.

What carries the argument

Place selection strategies that enforce a combined fitness guarantee together with a refined token-replay metric for isolated places.

If this is right

  • The produced nets can capture long-term dependencies while still meeting the stated fitness bound.
  • Precision stays high because only places that fit the chosen threshold are retained.
  • Infrequent labels are less likely to be excluded by the place evaluation step.
  • The method can be run on both real and artificial logs to produce comparable results across selection variants.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same selection logic could be transferred to other place-based miners that currently evaluate candidates independently.
  • Scaling the approach to very large logs may require additional pruning before the fitness checks are performed.
  • If the refined metric proves stable, it could replace the original metric in any replay-based discovery tool.

Load-bearing premise

Places that each replay enough traces in isolation can be assembled without creating deadlocks or dropping below the target fitness for the whole net.

What would settle it

An event log and noise threshold where the adapted algorithm returns a net that replays strictly less than the required fraction of traces.

Figures

Figures reproduced from arXiv: 2212.11047 by Lisa Luise Mannel, Wil M. P. van der Aalst.

Figure 1
Figure 1. Figure 1: The behavior in event log L corresponds in large parts to the sequential Petri net below. However, in all traces some deviations in activity order occur (marked in red). Since all traces and all activities are equally frequent, it is not possible to filter infrequent behavior patterns and discover the underlying main process structure by simply removing infrequent traces or activities in a preprocessing st… view at source ↗
Figure 2
Figure 2. Figure 2: Consider the event log L = [⟨▶, a, b, ■⟩ 40 ,⟨▶, b, a, ■⟩ 60], where the first trace variant occurs 40 times and the second one 60 times. Considered in isolation, place p6 allows for the first sequence of activities while place p7 allows for the second. However, in combination they cause a deadlock in the Petri net. In this paper, we aim to remedy the issue by investigating strategies of selecting a subset… view at source ↗
Figure 3
Figure 3. Figure 3: Petri net with A = {▶, a, b, c, ■} and P = {(▶|a, b),(▶|a, c),(a, b|■),(a, c|■),(∅|▶),(■|∅)}. The state of a Petri net N = (A, P) is defined by its marking. A marking M assigns tokens to places, i.e., M : P → N0, and can therefore be defined as a multiset of places. The Petri net in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Example of a tree-structured candidate space for the set of activities [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Illustrating the fitness status a candidate place can have with respect to an event log and noise [PITH_FULL_IMAGE:figures/full_fig_p013_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of the sets of places considered within the eST-Miner framework. To exemplify the [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: To exemplify the problems addressed in this work we use the Petri net above together with the event [PITH_FULL_IMAGE:figures/full_fig_p015_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Consider the event log [⟨▶, a, b, c, d, ■⟩ 35 ,⟨▶, a, b, c, e, ■⟩ 5 ,⟨▶, b, a, c, d, ■⟩ 55 ,⟨▶, b, a, c, e, ■⟩ 5 ]. The Petri net above is able to perfectly replay all traces in the event log. Alternatively, one can focus on the most frequent behavior and insert only p2, which would result in the removal of the then implicit places p1 and p6. This second option is illustrated in [PITH_FULL_IMAGE:figures/f… view at source ↗
Figure 9
Figure 9. Figure 9: Consider the event log [⟨▶, a, b, c, d, ■⟩ 35 ,⟨▶, a, b, c, e, ■⟩ 5 ,⟨▶, b, a, c, d, ■⟩ 55 ,⟨▶, b, a, c, e, ■⟩ 5 ]. The Petri net above (implicit places marked with gray, dashed lines) is able to replay the two trace variants ⟨▶, b, a, c, d, ■⟩ 55 and ⟨▶, b, a, c, e, ■⟩ 5 , which constitute a fraction of 0.6 of the input event log. To avoid the issue of deadlocks caused by combining places, in Section 6 we… view at source ↗
Figure 10
Figure 10. Figure 10: High-level overview of the proposed algorithmic framework. The newly introduced place selection [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Traces in L1 and L2 which are unfitting with respect to the place p = (a|b, c) are marked in red, while traces not activated are colored gray. For L1 we have that fmL1 agg(p) < fmL1 rel (p), while for L2 we have that fmL2 agg(p) > fmL2 rel (p). To obtain a metric that guarantees to satisfy a given fitness threshold τ with respect to all three fitness metrics discussed, we define combined fitness as a fina… view at source ↗
Figure 12
Figure 12. Figure 12: The table indicates for each of the given trace variants and candidate places whether the place can [PITH_FULL_IMAGE:figures/full_fig_p023_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Consider the set of places given in Figure 12. This figure shows all possible combinations of these [PITH_FULL_IMAGE:figures/full_fig_p023_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Overview of the presented approach, including input, output, and parameter use. [PITH_FULL_IMAGE:figures/full_fig_p025_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Example behavior of the delta adaption modifier [PITH_FULL_IMAGE:figures/full_fig_p028_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: An example model used to illustrate the applied quality metrics. [PITH_FULL_IMAGE:figures/full_fig_p030_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Overview of model quality results for all [PITH_FULL_IMAGE:figures/full_fig_p034_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Overview of quality results for all models discovered in the [PITH_FULL_IMAGE:figures/full_fig_p035_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Comparison of the number of dead transitions of the models discovered using [PITH_FULL_IMAGE:figures/full_fig_p036_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: The Petri nets discovered based on the Orders log using the Inductive Miner infrequent (default settings), the eST-Miner with τ = 1.0, and a subset of interesting models discovered using the presented approach [PITH_FULL_IMAGE:figures/full_fig_p037_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: The Petri nets discovered based on the Sepsis log using the Inductive Miner infrequent (default settings), the eST-Miner with τ = 1.0, and a subset of interesting models discovered using the presented approach [PITH_FULL_IMAGE:figures/full_fig_p038_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: The Petri nets discovered based on the RTFM log using the Inductive Miner infrequent (default settings), the eST-Miner with τ = 1.0, and a subset of interesting models discovered using the presented approach. model optionality of those activities. Additionally, the control-flow of the appeal-related subprocess is included. Here, self-loops are not only used to model skippable activities but also to enforc… view at source ↗
Figure 23
Figure 23. Figure 23: The Petri nets discovered based on the Teleclaims log using the Inductive Miner infrequent (default settings), the eST-Miner with τ = 1.0, and and a subset of interesting models discovered using the presented approach. by IMf express similar behavior, with the main difference being the representation of skippable activ￾ities: with all transitions being uniquely labeled, our approach has to rely on loop co… view at source ↗
Figure 24
Figure 24. Figure 24: Overview of the parameter choices resulting in the discovery of the models with the top [PITH_FULL_IMAGE:figures/full_fig_p043_24.png] view at source ↗
Figure 24
Figure 24. Figure 24: For the RTFM log, which has a few very dominant trace variants, we seem to generally achieve good results for rather high values of τ . In contrast, for the Sepsis log, which has a high variety of traces, a low τ -value seems mandatory to achieve high scores. Most likely, the large variety of fitting places allows for obtaining high precision, while our heuristics seems to successfully ensure the focus on… view at source ↗
Figure 25
Figure 25. Figure 25: Overview of the running times achieved in our experiments on the left, with a smaller scale visu [PITH_FULL_IMAGE:figures/full_fig_p045_25.png] view at source ↗
read the original abstract

In process discovery, the goal is to find, for a given event log, the model describing the underlying process. While process models can be represented in a variety of ways, Petri nets form a theoretically well-explored description language and are therefore often used. In this paper, we extend the eST-Miner process discovery algorithm. The eST-Miner computes a set of Petri net places which are considered to be fitting with respect to a certain fraction of the behavior described by the given event log as indicated by a given noise threshold. It evaluates all possible candidate places using token-based replay. The set of replayable traces is determined for each place in isolation, i.e., these sets do not need to be consistent. This allows the algorithm to abstract from infrequent behavioral patterns occurring only in some traces. However, when combining places into a Petri net by connecting them to the corresponding uniquely labeled transitions, the resulting net can replay exactly those traces from the event log that are allowed by the combination of all inserted places. Thus, inserting places one-by-one without considering their combined effect may result in deadlocks and low fitness of the Petri net. In this paper, we explore adaptions of the eST-Miner, that aim to select a subset of places such that the resulting Petri net guarantees a definable minimal fitness while maintaining high precision with respect to the input event log. Furthermore, current place evaluation techniques tend to block the execution of infrequent activity labels. Thus, a refined place fitness metric is introduced and thoroughly investigated. In our experiments we use real and artificial event logs to evaluate and compare the impact of the various place selection strategies and place fitness evaluation metrics on the returned Petri net.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper extends the eST-Miner algorithm, which discovers Petri net places via isolated token-based replay against a noise threshold. It introduces place-selection strategies to ensure the combined net achieves a user-specified minimum fitness while retaining high precision, and proposes a refined place fitness metric intended to avoid blocking infrequent activity labels. The adaptations and metric are evaluated on real and artificial event logs.

Significance. The work directly targets a known limitation of place-based miners: isolated fitness does not guarantee joint replayability. If the selection procedures deliver the claimed fitness bounds without excessive precision loss, the result would strengthen the practical utility of eST-Miner-style algorithms in noisy logs. The experimental comparison of multiple selection heuristics and the refined metric supplies concrete evidence on trade-offs.

major comments (2)
  1. [§4.2] §4.2 (place-selection procedure): the proof sketch that the greedy or exhaustive subset selection preserves the target fitness relies on the assumption that adding a place never decreases the replayable trace set; this needs an explicit invariant or counter-example check, because the token-replay semantics of the combined net can introduce deadlocks not visible in isolated places.
  2. [Table 2] Table 2 (fitness results on real logs): the reported fitness values for the refined metric are only 0.02–0.05 higher than the baseline, yet the paper claims the metric “thoroughly” prevents blocking of infrequent labels; the effect size appears too small to support the central claim without additional statistical tests or larger logs.
minor comments (2)
  1. [Eq. (7)] Notation for the refined fitness function (Eq. 7) re-uses the symbol f without re-defining it after the original token-replay fitness; a short clarifying sentence would avoid ambiguity.
  2. [§5] The experimental section lists log characteristics but omits the exact parameter settings (noise threshold, minimum place support) used for each run; these should be tabulated for reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment below and note the revisions that will be incorporated.

read point-by-point responses

  1. Referee: [§4.2] §4.2 (place-selection procedure): the proof sketch that the greedy or exhaustive subset selection preserves the target fitness relies on the assumption that adding a place never decreases the replayable trace set; this needs an explicit invariant or counter-example check, because the token-replay semantics of the combined net can introduce deadlocks not visible in isolated places.

    Authors: We agree that the proof sketch in §4.2 is insufficiently rigorous. Adding a place can only produce a non-increasing replayable trace set due to additional constraints, and deadlocks may arise from interactions not visible in isolation. The selection strategies aim to ensure the combined fitness bound by construction (via exhaustive enumeration or greedy addition with combined checks), but we will revise the section to state an explicit invariant: the selected subset is guaranteed to meet the target fitness because candidate places are only retained after verifying the joint replayability on the log. A formal argument or counter-example verification will be added. revision: yes

  2. Referee: [Table 2] Table 2 (fitness results on real logs): the reported fitness values for the refined metric are only 0.02–0.05 higher than the baseline, yet the paper claims the metric “thoroughly” prevents blocking of infrequent labels; the effect size appears too small to support the central claim without additional statistical tests or larger logs.

    Authors: The manuscript uses 'thoroughly investigated' to describe the evaluation of the metric, not a claim that it thoroughly prevents blocking. The modest aggregate fitness gains mask per-trace improvements where the refined metric avoids blocking infrequent labels that the baseline blocks. We will add statistical tests on the differences and results from additional larger synthetic logs to better substantiate the practical impact. revision: partial

Circularity Check

0 steps flagged

No significant circularity; claims rest on standard Petri net semantics and token replay

full rationale

The paper extends eST-Miner by selecting place subsets that guarantee definable minimal fitness in the combined net and introduces a refined place fitness metric. No equations, definitions, or self-citations reduce the fitness guarantees or metric to fitted parameters or prior results by construction. The approach uses independent place evaluation via token-based replay, with the combination step treated as a separate algorithmic adaptation rather than a self-referential fit. This is the expected non-finding for a paper whose central contribution is an engineering adaptation of established process-mining primitives.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper extends an established algorithm using standard domain concepts from process mining without introducing new free parameters, axioms beyond background Petri net theory, or invented entities.

axioms (1)
  • domain assumption Petri nets are a suitable formal language for process models and token-based replay can evaluate individual place fitness against an event log
    Invoked throughout the description of eST-Miner and its extensions as the evaluation mechanism.

pith-pipeline@v0.9.0 · 5847 in / 1185 out tokens · 33232 ms · 2026-05-24T10:36:28.037333+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages

  1. [1]

    Understanding Petri Nets: Modeling Techniques, Analysis Methods, Case Studies

    Reisig W. Understanding Petri Nets: Modeling Techniques, Analysis Methods, Case Studies. Springer Berlin Heidelberg, 2013. ISBN:9783642332784

    work page 2013

  2. [2]

    Petri nets and business process management

    Desel J, Oberweis A, Reisig W, Rozenberg G. Petri nets and business process management. Saarbr¨ucken: Gesch¨aftsstelle Schloss Dagstuhl, 1998

    work page 1998

  3. [3]

    Free-Choice Petri Nets

    Desel J, Esparza J. Free-Choice Petri Nets. Cambridge Tracts in Theoretical Computer Science. Cambridge University Press, 1995. ISBN:9780521465199

    work page 1995

  4. [4]

    Transformations and Decompositions of Nets

    Berthelot G. Transformations and Decompositions of Nets. In: Petri Nets: Central Models and Their Properties. Springer, Berlin, Heidelberg, 1987 pp. 359–376

    work page 1987

  5. [5]

    Process Mining: Data Science in Action

    van der Aalst W. Process Mining: Data Science in Action. Springer, Heidelberg, 2 edition, 2016. doi:10.1007/978-3-662-49851-4

    work page doi:10.1007/978-3-662-49851-4 2016

  6. [6]

    Mining process models with non-free-choice constructs

    Wen L, van der Aalst WMP, Wang J, Sun J. Mining process models with non-free-choice constructs. Data Mining and Knowledge Discovery, 2007. 15(2):145–180. doi:10.1007/s10618-007-0065-y. L.L. Mannel and W. M.P . van der Aalst/ Discovering Process Models while Providing Guarantees... 157

    work page doi:10.1007/s10618-007-0065-y 2007

  7. [7]

    Discovering Block-Structured Process Models from Event Logs - A Constructive Approach

    Leemans S, Fahland D, van der Aalst W. Discovering Block-Structured Process Models from Event Logs - A Constructive Approach. Application and Theory of Petri Nets and Concurrency, 2013. Lecture Notes in Computer Science, vol 7927. doi:10.1007/978-3-642-38697-8 17

    work page doi:10.1007/978-3-642-38697-8 2013

  8. [8]

    Flexible Heuristics Miner (FHM)

    Weijters AJMM, Ribeiro JTS. Flexible Heuristics Miner (FHM). In: Proceedings of the IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2011, part of the IEEE Symposium Series on Computational Intelligence 2011, April 11-15, 2011, Paris, France. IEEE, 2011 pp. 310–317. doi:10.1109/CIDM.2011.5949453

    work page doi:10.1109/cidm.2011.5949453 2011

  9. [9]

    Petri Net Synthesis

    Badouel E, Bernardinello L, Darondeau P. Petri Net Synthesis. Text in Theoretical Computer Science, EATCS Series. Springer, 2015. doi:10.1007/978-3-662-47967-4

    work page doi:10.1007/978-3-662-47967-4 2015

  10. [10]

    How to Synthesize Nets from Languages: A Survey

    Lorenz R, Mauser S, Juh ´as G. How to Synthesize Nets from Languages: A Survey. In: Proceedings of the 39th Conference on Winter Simulation: 40 Years! The Best is Yet to Come, WSC ’07. IEEE Press, Piscataway, NJ, USA, 2007 pp. 637–647. doi:10.1109/WSC.2007.4419657

    work page doi:10.1109/wsc.2007.4419657 2007

  11. [12]

    Process Discovery Using Integer Linear Programming

    van der Werf JM, van Dongen B, Hurkens C, Serebrenik A. Process Discovery Using Integer Linear Programming. In: Applications and Theory of Petri Nets. Springer, Berlin, Heidelberg, 2008. doi:10.1007/978-3-540-68746-7 24

    work page doi:10.1007/978-3-540-68746-7 2008

  12. [13]

    Avoiding Over-Fitting in ILP-Based Process Discovery

    van Zelst S, van Dongen B, van der Aalst W. Avoiding Over-Fitting in ILP-Based Process Discovery. In: Business Process Management. Springer International Publishing, Cham, 2015 pp. 163–171. doi:10.1007/978-3-319-23063-4 10

    work page doi:10.1007/978-3-319-23063-4 2015

  13. [14]

    ILP-Based Process Discovery Using Hybrid Regions

    van Zelst S, van Dongen B, van der Aalst W. ILP-Based Process Discovery Using Hybrid Regions. In: ATAED@Petri Nets/ACSD. 2015

    work page 2015

  14. [15]

    A region-based algorithm for discovering Petri nets from event logs

    Carmona J, Cortadella J, Kishinevsky M. A region-based algorithm for discovering Petri nets from event logs. In: Business Process Management. Springer, 2008 p. 358–373

    work page 2008

  15. [16]

    Deriving unbounded Petri nets from formal languages

    Darondeau P. Deriving unbounded Petri nets from formal languages. In: CONCUR’98 Concurrency Theory. Springer, Berlin, Heidelberg. 1998 pp. 533–548. ISBN:978-3-540-68455-8

    work page 1998

  16. [17]

    Synthesis of Petri Nets from Finite Partial Languages

    Bergenthum R, Desel J, Lorenz R, Mauser S. Synthesis of Petri Nets from Finite Partial Languages. Fundam. Informaticae, 2008. 88(4):437–468

    work page 2008

  17. [18]

    Partial (set) 2-structures

    Ehrenfeucht A, Rozenberg G. Partial (set) 2-structures. Acta Informatica, 1990. 27(4):343–368

    work page 1990

  18. [19]

    A Symbolic Algorithm for the Synthesis of Bounded Petri Nets

    Carmona J, Cortadella J, Kishinevsky M, Kondratyev A, Lavagno L, Yakovlev A. A Symbolic Algorithm for the Synthesis of Bounded Petri Nets. In: van Hee KM, Valk R (eds.), Applications and Theory of Petri Nets. Springer Berlin Heidelberg, Berlin, Heidelberg. 2008 pp. 92–111. ISBN:978-3-540-68746-7

    work page 2008

  19. [20]

    Automated Repair of Process Models Using Non-Local Constraints

    Kalenkova A, Carmona J, Polyvyanyy A, La Rosa M. Automated Repair of Process Models Using Non-Local Constraints. In: Application and Theory of Petri Nets and Concurrency. Springer-Verlag, Berlin, Heidelberg. 2020 p. 280–300. ISBN:978-3-030-51830-1

    work page 2020

  20. [21]

    Finding Complex Process-Structures by Exploiting the Token-Game

    Mannel LL, van der Aalst WMP. Finding Complex Process-Structures by Exploiting the Token-Game. In: Donatelli S, Haar S (eds.), Application and Theory of Petri Nets and Concurrency - 40th International Conference, PETRI NETS 2019, Aachen, Germany, June 23-28, 2019, Proceedings, volume 11522 of Lecture Notes in Computer Science. Springer, 2019 pp. 258–278. ...

    work page doi:10.1007/978-3-030-21571-2 2019

  21. [22]

    Waiting Nets

    Mannel LL, van der Aalst WMP. Discovering Process Models with Long-Term Dependencies While Providing Guarantees and Handling Infrequent Behavior. In: Bernardinello L, Petrucci L (eds.), Application and Theory of Petri Nets and Concurrency - 43rd International Conference, PETRI NETS 2022, Bergen, Norway, June 19-24, 2022, Proceedings, volume 13288 of Lectu...

    work page doi:10.1007/978-3-031-06653-5 2022

  22. [23]

    Implicit places in net systems

    Garcia-Valles F, Colom J. Implicit places in net systems. Proceedings 8th International Workshop on Petri Nets and Performance Models, 1999. pp. 104–113. doi:10.1109/PNPM.1999.796557

    work page doi:10.1109/pnpm.1999.796557 1999

  23. [24]

    Petri Net Reductions for Counting Markings

    Berthomieu B, Botlan DL, Dal-Zilio S. Petri Net Reductions for Counting Markings. In: Gallardo M, Merino P (eds.), Model Checking Software - 25th International Symposium, SPIN 2018, Malaga, Spain, June 20-22, 2018, Proceedings, volume 10869 of LNCS. Springer, 2018 pp. 65–84

    work page 2018

  24. [25]

    Improving the linearly based characterization of P/T nets

    Colom J, Silva M. Improving the linearly based characterization of P/T nets. In: Advances in Petri Nets

    work page

  25. [26]

    Springer, Berlin, Heidelberg, 1991 pp. 113–145. doi:10.1007/3-540-53863-1 23

    work page doi:10.1007/3-540-53863-1 1991

  26. [27]

    Removing Implicit Places Using Regions for Process Discovery

    Mannel LL, Bergenthum R, van der Aalst WMP. Removing Implicit Places Using Regions for Process Discovery. In: Proceedings of the International Workshop on Algorithms & Theories for the Analysis of Event Data (ATAED) 2020, volume 2625. CEUR-WS.org pp. 20–32

    work page 2020

  27. [28]

    Discovering the ”Glue” Connecting Activities - Exploiting Monotonicity to Learn Places Faster

    van der Aalst WMP. Discovering the ”Glue” Connecting Activities - Exploiting Monotonicity to Learn Places Faster. In: It’s All About Coordination - Essays to Celebrate the Lifelong Scientific Achievements of Farhad Arbab. 2018 pp. 1–20. doi:10.1007/978-3-319-90089-6 1

    work page doi:10.1007/978-3-319-90089-6 2018

  28. [29]

    Conformance Checking - Relating Processes and Models

    Carmona J, van Dongen B, Solti A, Weidlich M. Conformance Checking - Relating Processes and Models. Springer, Cham, 2018. doi:10.1007/978-3-319-99414-7

    work page doi:10.1007/978-3-319-99414-7 2018

  29. [30]

    Sepsis Cases - Event Log, 2016

    Mannhardt, F. Sepsis Cases - Event Log, 2016. doi:10.4121/UUID:915D2BFB-7E84-49AD-A286-DC35F063A460

    work page doi:10.4121/uuid:915d2bfb-7e84-49ad-a286-dc35f063a460 2016

  30. [31]

    Road Traffic Fine Management Process, 2015

    De Leoni, M, Mannhardt, F. Road Traffic Fine Management Process, 2015

    work page 2015

  31. [32]

    Spreadsheets for BPM

    van der Aalst WMP. Spreadsheets for BPM. Business Process Management Journal, 2010. 24:105–127

    work page 2010

  32. [33]

    Aligning observed and modeled behavior

    Adriansyah A. Aligning observed and modeled behavior. Ph.D. thesis, Mathematics and Computer Science, 2014. doi:10.6100/IR770080

    work page doi:10.6100/ir770080 2014

  33. [34]

    A Fresh Look at Precision in Process Conformance

    Munoz-Gama J, Carmona J. A Fresh Look at Precision in Process Conformance. In: BPM, volume 6336. 2010 pp. 211–226. ISBN:978-3-642-15617-5

    work page 2010

  34. [35]

    Soundness of Resource-Constrained Workflow Nets

    van Dongen B, de Medeiros A, Verbeek H, Weijters A, van der Aalst W. The ProM Framework: A New Era in Process Mining Tool Support. In: Applications and Theory of Petri Nets 2005. Springer, Berlin, Heidelberg, 2005 pp. 444–454. doi:10.1007/11494744 25

    work page doi:10.1007/11494744 2005

  35. [36]

    What Makes Process Models Understandable? In: Alonso G, Dadam P, Rosemann M (eds.), Business Process Management

    Mendling J, Reijers HA, Cardoso J. What Makes Process Models Understandable? In: Alonso G, Dadam P, Rosemann M (eds.), Business Process Management. Springer Berlin Heidelberg, 2007 pp. 48–63. doi:10.1007/978-3-540-75183-0 4

    work page doi:10.1007/978-3-540-75183-0 2007

  36. [37]

    Discovering more precise process models from event logs by filtering out chaotic activities

    Tax N, Sidorova N, van der Aalst WMP. Discovering more precise process models from event logs by filtering out chaotic activities. J. Intell. Inf. Syst., 2019. 52(1):107–139. doi:10.1007/s10844-018-0507-6

    work page doi:10.1007/s10844-018-0507-6 2019