Pith.

Databases · arXiv:2608.09398 v1 · submitted 10 Aug 2026 · 46 references

Monotonicity-Guided Bottom-Up Petri Net Discovery: The SPECpp Framework

Leah Tacke genannt Unterberg, Lisa L. Mannel, Wil M. P. van der Aalst

The paper says

“Process discovery is one of the central challenges in process mining. Petri nets are particularly attractive because simple local constructs can express complex behavior, including concurrency.”

more of the abstract

While their global behavior may be difficult to analyze, individual places can be efficiently characterized using monotonic properties, enabling bottom-up discovery. Unlike top-down approaches such as the Inductive Miner, which rely on predefined constructs for sequences, choices, loops, and concurrency, our approach allows such structures to emerge organically and can exploit the full expressive power of Petri nets, including free-choice constructs and long-term dependencies. The main challenge is the exponential number of candidate places and their combinations. We present the SPECpp framework which implements strategies to obtain high-quality models under time and resource constraints. SPECpp supports rapid experimentation and is used to evaluate these strategies using both synthetic and real-life event data.

from the abstract

Pith says

Machine review · deepseek-v4-flash · 11 Aug 2026 · what was read was not recorded · not checked by a person

It can be wrong. Challenge this review or reply to it.

Holds up, with conditions. A real framework contribution with solid formalization and released code, but the headline completeness guarantee is conditional in a way the implementation itself can violate, and the evaluation lacks baselines. This line is wrong

3 major objections · 3 minor

How often objections from reviews like this one are right: not yet measured

Are you an author? What you can do.

Not checked: figures, code, what the 46 references say.

The paper says, at §5 (Tables 3–4, Figures 8–15)

The review did not quote this passage.

Pith says: Objection 1 of 3 · major

The evaluation contains no quantitative comparison with other discovery algorithms on the same logs. The motivating example in Figure 2 contrasts eST-Miner, Alpha Miner, Inductive Miner, and ILP Miner qualitatively, but the experiments report only SPECpp's own fitness, precision, and number of places.

more of objection 1

Consequently, phrases such as 'high-quality models' and 'best scoring models' are not calibrated against any baseline, and the conclusion that the runtime is 'not yet competitive' is not supported by measurements. I recommend adding baselines, including the underlying eST-Miner variants, evaluated with identical metrics and time limits, or substantially softening the evaluative claims to feasibility demonstrations.

The paper says, at §3.1, §3.4.1, §4.1

The review did not quote this passage.

Pith says: Objection 2 of 3 · major

Proposition 1's completeness and minimality guarantee is conditional on assumptions that are not verified for the evaluated configurations. First, the not-biwiring constraint in §3.4.1 is explicitly stated to be able to exclude previously proposed candidates, which violates the minimality assumption of Proposition 1.

more of objection 2

Second, §4.1 states that the implementation permits non-subtree-monotonic constraints and deletion of wiring constraints at the cost of incomplete traversal. Third, the proof of Proposition 1 assumes termination by exhaustion, but Table 3 shows many runs stopped by cooperative time-out (e.g., 56.7% PEC timeouts for BPIC12), and Table 2's depth-limited candidate trees are not complete over all possible places. The paper does not report which of the 60 runs satisfy the conditions needed for Proposition 1, so the reported models cannot be claimed to inherit the complete-and-minimal traversal guarantee. Please either restrict the formal claim to runs that exhaust a complete tree under monotone non-excluding updates, or present the evaluated mode as heuristic and support its quality by other evidence.

The paper says, at §5 (Figures 8–15)

The review did not quote this passage.

Pith says: Objection 3 of 3 · major

The model-selection protocol is optimistic. Figures 8–15 show the top 25% F1 models chosen after inspecting all 60 parameter combinations on the same logs for which fitness and precision are reported, with no holdout or separate evaluation set. The many DNF runs summarized in Table 3 are excluded from this selection, which further biases the reported statistics toward favorable outcomes.

more of objection 3

I recommend reporting all configurations or a clearly pre-specified subset, and separating configuration search from evaluation, or at least labeling the figures explicitly as best-case demonstrations rather than expected performance.

3 minor objections

The paper says, at §3.4.1 (Lemma 3)

Pith says: minor objection 1 of 3

Lemma 3 contains an unbound variable p′ in the inequality; the statement should read 'for every place p′ in the union ... #△_L,p ≤ #△_L,p′'.

The paper says, at §5 (Figure 13)

Pith says: minor objection 2 of 3

The captions in Figure 13 say 'Spearman correlation efficient' instead of 'coefficient'; please correct the typo.

show 1 more

The paper says, at §3.1 (Definition 13)

Pith says: minor objection 3 of 3

Definition 13 permits a monotonic strengthening to exclude a previously returned place, which is a different condition from the 'never excludes previously returned places' assumption used in Proposition 1. The distinction is important and should be stated explicitly where Proposition 1 is introduced.

What would settle it

Use the released implementation to run the uniwired configuration on a minimal log containing a trace such as ⟨▶,a,c,b,d,■⟩, with τ=1.0, max tree depth 4, and ReplaceExisting enabled. Instrument `propose` to record the returned place set M and the final constraint set C_T, and check whether M equals P↾I_C_T. Specifically test whether q=({a},{b}) is returned and later excluded by the not-biwiring constraint generated when p=({a,c},{b,d}) replaces q. If this occurs, the minimality half of Proposition 1 fails for that configuration; if the implementation avoids it by restarting traversal or by never generating such a constraint, report the exact condition under which the guarantee holds.

from the skeptic's reading

How the review went

4 readings by deepseek-v4-flash · 11 Aug 2026

  1. Reader holds up, with conditions confidence high
  2. Skeptic agrees 4 min 1 s · 28,224 tokens read, 20,593 written
  3. Editor wrote the verdict above 29 s · 28,990 tokens read, 2,900 written
  4. Referee major revision: 3 major objections, 3 minor 2 min 22 s · 28,551 tokens read, 13,229 written

Not run on this paper: the simulated authors' defence.

Pith did not check this review's objections against the paper.

The review in full

The review in brief, machine-written

This is a genuine framework paper: the PEC cycle, the efficient candidate tree with tree-state traversal, and the cleaned-up formalization of the eST-Miner line are new and presented clearly. The proofs of Theorem 1, Proposition 1, and Theorem 3 look correct under their stated assumptions, and the authors ship real code plus a ProM plugin, which makes the framework reusable. The evaluation is honest about timeouts and limitations, and the write-up does not pretend the current runtime is competitive.

The main soft spot is the gap between the headline guarantee and the implementation. Proposition 1 promises complete and minimal candidate traversal, but only under subtree-monotonic constraints that are strengthened monotonically and never exclude previously returned candidates. The paper's own not-biwiring constraint (Section 3.4.1) explicitly excludes previously returned places, and Section 4.1 admits the implementation can drop subtree-monotonicity, which voids the guarantee. The evaluation does not state which of the 60 runs honor the assumptions, so we do not actually know when the returned model enjoys the complete-and-minimal property. This is not fatal — a framework can offer formal guarantees for one configuration and heuristics for another — but the abstract overstates the expressiveness and the guarantees.

Two further weaknesses worth naming. There is no quantitative comparison against other discovery algorithms on the same logs, so it is hard to know how the models compare to Inductive Miner or ILP-based approaches. And the abstract claims free-choice constructs and long-term dependencies as though they come for free, while the implementation only supports unique labels, no silent transitions, and open-world semantics; those limitations are acknowledged in the body but not the abstract.

On the credit side, the self-citation is not a problem here — this line of work is the direct antecedent, and the framework genuinely generalizes it. The formal statements are careful about their assumptions, which is more than many framework papers do.

Who should read it: anyone working on bottom-up process discovery or extending eST-Miner. It deserves a serious referee; with revisions to the abstract, added baselines, and explicit documentation of when completeness is lost, it would be a solid platform contribution.

I'd engage with it.

The strongest objection

The central formal claim is Proposition 1: the proposal routine returns exactly the places satisfying the final constraint set under monotonic strengthenings plus the additional non-exclusion assumption. That additional assumption is not implied by Definition 13, and the paper's own constraint-generation example in §3.4.1 violates it. If q=({a},{b}) is returned first and p=({a,c},{b,d}) is later accepted, the not-biwiring constraint for p excludes q because q shares preset element a and postset element b with p, even though q was already returned. Thus M is not a subset of P↾I_C_T, so minimality fails even when all constraints are subtree-monotonic. Section 4.1 then states that the implementation 'can permit non subtree-monotonic constraints at the cost of incomplete traversal', and that wiring constraints can be deleted after a place is retroactively removed, which can also break completeness by never revisiting pruned subtrees. The evaluation does not state which of the 60 runs respect the assumptions needed for Proposition 1, so the reported models are not shown to enjoy the complete-and-minimal traversal guarantee.

The reader's weakest assumption identifies the same load-bearing point: the completeness and minimality guarantee depends on subtree-monotonic constraints and on updates never excluding previously returned candidates, while Section 4.1 admits the implementation can relax these conditions. My read strengthens this by noting that even the conceptual framework's uniwired constraint generator, not just an implementation contingency, can exclude a previously returned ancestor, so the minimality claim is narrower than the 'complete and minimal traversal' language suggests. The formal Proposition 1 is sound under its explicit assumptions, and the paper is honest about the gap, so this is not a rejection-level flaw. The paper also has real independent support: released code, a ProM plugin, and an evaluation that reports timeouts and non-competitive runtime. The conditional verdict remains appropriate, but the paper should state clearly that the guarantee applies only to configurations satisfying non-exclusion, and should identify which evaluation runs satisfy it. The proposed concrete test settles whether the uniwired configuration violates the guarantee in practice and whether the implementation's restart option is needed for completeness.

The referee's report

If the formal traversal guarantee is usable in practice, the framework is a valuable contribution: it gives a general, extensible architecture for bottom-up discovery, states explicit conditions under which candidate traversal is complete and minimal, and ships with open-source code and a ProM plugin. The paper is also unusually transparent about its limitations, including timeouts, the heuristic character of some components, and the gap between the conceptual framework and implemented constraint handling. However, the empirical evaluation does not calibrate the reported quality against other discovery algorithms, and it does not connect the formal guarantee to the configurations actually run in Section 5. The paper is therefore best read as a systems/formalization contribution whose empirical claims need additional support before the framework can be positioned as delivering high-quality models in practice.

Pith's reading of the paper (a machine's summary, not the authors')

This paper sets out to show that bottom-up Petri net discovery can be made practical by exploiting monotone properties of places. It presents SPECpp, a Proposal–Evaluation–Composition loop with post-processing, in which candidate places are generated from a tree whose subtree relation matches monotone constraints, so that a single local evaluation can prune a whole family of candidates. The formal payload is Proposition 1: if constraint updates are monotonic strengthenings that never exclude an already returned place, the proposal routine returns exactly the set of places satisfying the final constraints. If this holds, bottom-up discovery can find long-term dependencies and non-free-choice behavior that block-structured top-down miners miss, without paying for the exponential candidate space. The paper also reports an implementation and an evaluation on synthetic and real event logs showing that high-quality models are often found quickly, though full traversals can time out on complex logs.

What carries the argument. The machinery is an efficient candidate tree: a rooted tree over all possible places, generated on demand by a child function g with unique parents, where each child is an incremental preset or postset expansion of its parent. Constraints are subtree-monotonic if a constraint that fails at a place also fails at every descendant; this turns one local evaluation into a whole-subtree prune. The paper's central identity is the equality $\mathrm{cand}_g(p) \restriction I_C = \mathrm{cand}_{g_C}(p)$ for subtree-monotonic $C$ when $I_C(p)$ holds, and the under/overfedness monotonicity that makes fitness a valid source of such subtree-monotonic constraints. The monotonic-strengthening requirement is what upgrades pruning from a heuristic to a completeness and minimality guarantee.

Core claim. The central claim is that bottom-up Petri net discovery can be organized as a Proposal–Evaluation–Composition loop in which the exponential space of candidate places is represented implicitly by a candidate tree, and monotone properties of places make that tree prunable without losing solutions. Specifically, Proposition 1 claims that if constraint sets are subtree-monotonic and are updated only by monotonic strengthening that never excludes an already returned candidate, the proposal routine returns exactly the set of places satisfying the final constraints. The paper argues this makes bottom-up discovery expressive enough to capture long-term dependencies and non-free-choice behavior that block-structured top-down miners miss, and the accompanying implementation demonstrates the strategies on synthetic and real event logs.

The premise it rests on. The completeness and minimality guarantee assumes constraints are subtree-monotonic and only ever strengthen without excluding already returned candidates; the implementation permits non-subtree-monotonic constraints, such as deleting a wiring constraint after a place is removed, and in that mode candidate places can silently be lost.

If the paper is right (Pith's inferences, not the authors')

A discovery run that only strengthens constraints is guaranteed not to miss any place that meets the final constraints, so the resulting set of places can be trusted as complete with respect to those constraints.

Because the model is assembled from individual places rather than predefined blocks, constructs like long-term dependencies and non-free-choice behavior can appear whenever the log supports them, which is exactly the expressiveness gap of top-down miners.

One local evaluation of a place can rule out an entire subtree of the candidate tree, which is how the exponential number of activity combinations is tamed without exhaustive enumeration.

Intermediate results stay small through relative evaluations such as implicit-place removal, so more expensive model-level checks can be applied on the composition side.

The evaluation runs indicate that good models are often found long before the candidate space is exhausted, meaning ordering and pruning strategies can make deliberately incomplete runs practically useful.

Implicit in the paper: the same Proposal–Evaluation–Composition loop should generalize to any locally checkable, tree-monotone property, not just fitness; partial cuts or object-centric dependencies could be proposed and pruned the same way.

A testable extension would run the same log under a strictly monotone constraint policy and under the implementation's deletable wiring constraints, then compare the returned place sets to quantify how often the shipped heuristic mode loses candidates.

The framework's completeness argument is about places, not models; extending tree-monotone constraints to sets of places would be the natural next step toward guarantees on the final Petri net rather than on the candidate set.

On noisy real logs the inversion between the fitness threshold and global model fitness suggests that an adaptive per-subtree threshold, rather than a single fixed threshold, could improve both speed and quality; the paper leaves this unexplored.

Check this review

Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.

How often reviews like this one are right

Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.

How often objections from reviews like this one are right: not yet measured

Every value above, in the review's public record (JSON)

Reviews by people

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review

How Pith checks papers · The full audit of this paper

The paper

Abstract

Process discovery is one of the central challenges in process mining. Petri nets are particularly attractive because simple local constructs can express complex behavior, including concurrency. While their global behavior may be difficult to analyze, individual places can be efficiently characterized using monotonic properties, enabling bottom-up discovery. Unlike top-down approaches such as the Inductive Miner, which rely on predefined constructs for sequences, choices, loops, and concurrency, our approach allows such structures to emerge organically and can exploit the full expressive power of Petri nets, including free-choice constructs and long-term dependencies. The main challenge is the exponential number of candidate places and their combinations. We present the SPECpp framework which implements strategies to obtain high-quality models under time and resource constraints. SPECpp supports rapid experimentation and is used to evaluate these strategies using both synthetic and real-life event data.

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.

abstract compiled · 2026-09-23

The authors' main claim, as the classifier picked it out 0.68

“We present the SPECpp framework which implements strategies to obtain high-quality models under time and resource constraints.”

How the main claim is stated: asserted 1.00

What the abstract itself states, as probabilities: claims first 0.05 · states limitation 0.68 · human participants 0.05 · controlled comparison 0.17 · randomized assignment 0.01 · states quantitative result 0.04 · compares with prior results 0.81 · mentions code or data availability 0.05

The abstract, sentence by sentence

This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.

SentenceRoleThe authors' own workStates the main conclusionStrength
Process discovery is one of the central challenges in process mining. background 1.00 0.04 0.02 asserted 1.00
Petri nets are particularly attractive because simple local constructs can express complex behavior, including concurrency. background 1.00 0.05 0.03 asserted 0.99
While their global behavior may be difficult to analyze, individual places can be efficiently characterized using monotonic properties, enabling bottom-up discovery. background 0.66 0.14 0.13 asserted 0.88
Unlike top-down approaches such as the Inductive Miner, which rely on predefined constructs for sequences, choices, loops, and concurrency, our approach allows such structures to emerge organically and can exploit the full expressive power of Petri nets, including free-choice constructs and long-term dependencies. contribution 0.89 0.50 0.19 asserted 0.97
The main challenge is the exponential number of candidate places and their combinations. background 0.49 0.24 0.03 asserted 1.00
We present the SPECpp framework which implements strategies to obtain high-quality models under time and resource constraints. contribution 0.98 0.88 0.22 asserted 1.00
SPECpp supports rapid experimentation and is used to evaluate these strategies using both synthetic and real-life event data. method 0.76 0.87 0.08 asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from observation or experiment 0.69

What was examined, and what was not

Sources read

  • arXiv listing abstract v1

Not examined

  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers

Levels run

  • P abstract

Not run

  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

How the labels were made

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 5,466 input tokens.

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON

Figures (15)
Figure 1 of the paper
Figure 1. A high-level overview of the framework flow. An event log is taken
Figure 2 of the paper
Figure 2. Petri nets discovered by various algorithms on the input log
Figure 3 of the paper
Figure 3. An exemplary partial efficient candidate tree Tex = (g, r) rooted at place r = ({a}, {}). all constraints collected up to that moment. Constraints are how candidate evaluations and their indirect impact are incorporated into the candidate traversal. Definition 9 (Constraints). A constraint c : P → {True, False} is a predi￾cate function over the set of possible places. A place p ∈ P meets constraint c, if and only if c(p) = True. To lift this to sets of constraints C, we use an indicator function IC(p) :⇔ ∀c ∈ C : c(p) = True. We use it to filter sets of places S ⊆ P to those meeting
Figure 4 of the paper
Figure 4. A schematic of how the subtree-monotonicity of a constraint set
Figure 5 of the paper
Figure 5. An example of the candidate subtree rooted at
Figure 6 of the paper
Figure 6. An overview of the main PEC components, their dependencies
Figure 7 of the paper
Figure 7. The discovery view of the ProM plugin. The left split shows the currently accepted places in a list view.
Figure 8 of the paper
Figure 8. The distribution and Pareto front of fitness vs. precision of the
Figure 9 of the paper
Figure 9. The distribution and Pareto front of fitness vs. #places (lower is
Figure 10 of the paper
Figure 10. The distribution and Pareto front of precision vs. #places (lower
Figure 11 of the paper
Figure 11. The percentile rank of the fastest execution to reach a model scor
Figure 12 of the paper
Figure 12. The average minimal max tree depth at which the model quality metrics are already maximized at. The average is over the different values of τ , as it has a big impact on the achievable metric. We see that, especially on the smaller synthetic logs, a high model fitness can be achieved with very low tree depth, thus saving running time. Precision appears to require more tree levels on average on these logs, and particularly to maximize both at the same time (F1-score), more candidate places are required. 46
Figure 13 of the paper
Figure 13. The correlation between the varied parameters and model metrics.
Figure 14 of the paper
Figure 14. One of the best models for Teleclaims, discovered for τ = 0.8 and depths 4 and 5. It contains many long-term dependencies, e.g., between determine likelihood of claim and initiate payment. Its scores are 0.91 for fitness and 1 for precision. 48
Figure 15 of the paper
Figure 15. One of the best models for the RTFM log. It was discovered on three different runs with τ = 0.7 in 1.8s (depth limit 3), 10.4s (depth limit 4) and 45.3s (depth limit 5). It has a fitness of 0.93, precision of 1 and perfectly fits 68% of the log. 49

Figures are the authors', reproduced from the paper.

References (46)
  1. [1] van der Aalst, W.M.P.: Process Mining: Data Science in Action. Springer, Berlin, Heidelberg (2016) on Pith
  2. [2] van der Aalst, W.M.P.: Discovering the “glue” connecting activities - exploiting monotonicity to learn places faster. In: F.S. de Boer, M.M. Bonsangue, J. Rutten (eds.) It’s All About Coordination - Essays to Celebrate the Lifelong Scientific Achievements of Farhad Arbab,Lectu... on Pith
  3. [3] van der Aalst, W.M.P.: A Practitioner’s Guide to Process Mining: Lim- itations of the Directly-Follows Graph. In: International Conference on Enterprise Information Systems (Centeris 2019),Procedia Computer Science, vol. 164, pp. 321–328. Elsevier (2019) on Pith
  4. [4] van der Aalst, W.M.P.: Object-centric process mining: Dealing with divergence and convergence in event data. In: P.C. Olveczky, G. Salaün (eds.) Software Engineering and Formal Methods - 17th International Conference, SEFM2019,Lecture Notes in Computer Science, vol.11724, pp. ... on Pith
  5. [5] van der Aalst, W.M.P.: Discovering directly-follows complete Petri nets from event data. In: A Journey from Process Algebra via Timed Au- tomata to Model Learning,Lecture Notes in Computer Science, vol. 13560, pp. 539–558. Springer (2022) on Pith
  6. [6] van der Aalst, W.M.P.: Foundations of Process Discovery. In: W.M.P. van der Aalst, J. Carmona (eds.) Process Mining Handbook,Lecture Notes in Business Information Processing, vol. 448, pp. 37–75. Springer (2022) on Pith
  7. [7] van der Aalst, W.M.P., Carmona, J. (eds.): Process Mining Handbook, Lecture Notes in Business Information Processing, vol. 448. Springer (2022) on Pith
  8. [8] van der Aalst, W.M.P., van Dongen, B.F.: Discovering Petri nets from eventlogs. TransactionsonPetriNetsandOtherModelsofConcurrency 7, 372–422 (2013) on Pith
  9. [9] van der Aalst, W.M.P., van Dongen, B.F., Herbst, J., Maruster, L., Schimm, G., Weijters, A.J.M.M.: Workflow mining: A survey of issues and approaches. Data Knowl. Eng.47(2), 237–267 (2003) 54 on Pith
  10. [10] van der Aalst, W.M.P., de Medeiros, A.K.A., Weijters, A.J.M.M.: Ge- netic process mining. In: G. Ciardo, P. Darondeau (eds.) Applications and Theory of Petri Nets 2005, 26th International Conference, ICATPN 2005,Lecture Notes in Computer Science, vol. 3536, pp. 48–69. Springer (2005) on Pith
  11. [11] van der Aalst, W.M.P., Rubin, V.A., Verbeek, H.M.W., van Dongen, B.F., Kindler, E., Günther, C.W.: Process mining: A two-step approach to balance between underfitting and overfitting. Softw. Syst. Model. 9(1), 87–111 (2010) on Pith
  12. [12] van der Aalst, W.M.P., Weijters, A.J.M.M., Maruster, L.: Workflow mining: Discovering process models from event logs. IEEE Trans. Knowl. Data Eng.16(9), 1128–1142 (2004) on Pith
  13. [13] Augusto, A., Carmona, J., Verbeek, E.: Advanced Process Discovery Techniques. In: W.M.P. van der Aalst, J. Carmona (eds.) Process Min- ing Handbook,Lecture Notes in Business Information Processing, vol. 448, pp. 76–107. Springer (2022) on Pith
  14. [14] Augusto, A., Conforti, R., Dumas, M., La Rosa, M., Polyvyanyy, A.: Split miner: Automated discovery of accurate and simple business pro- cess models from event logs. Knowl. Inf. Syst.59(2), 251–284 (2019) on Pith
  15. [15] Badouel, É., Bernardinello, L., Darondeau, P.: Petri Net Synthesis. Texts in Theoretical Computer Science. An EATCS Series. Springer (2015) on Pith
  16. [16] Brons, D., Scheepens, R., Fahland, D.: Striking a new balance in ac- curacy and simplicity with the probabilistic inductive miner. CoRR abs/2109.06288(2021). URLhttps://arxiv.org/abs/2109.06288 on PithPith's reviewarXiv
  17. [17] Buijs, J.C.A.M., van Dongen, B.F., van der Aalst, W.M.P.: On the role of fitness, precision, generalization and simplicity in process discovery. In: R. Meersman, et al. (eds.) On the Move to Meaningful Internet Systems: OTM 2012 - Confederated International Conferences: CoopIS... on Pith
  18. [18] Carmona, J., Cortadella, J., Kishinevsky, M.: A region-based algorithm for discovering Petri nets from event logs. In: M. Dumas, M. Reichert, 55 M. Shan (eds.) Business Process Management, 6th International Con- ference, BPM 2008,Lecture Notes in Computer Science, vol. 5240, p... on Pith
  19. [19] van Eck, M.L., Buijs, J.C.A.M., van Dongen, B.F.: Genetic process mining: Alignment-based process model mutation. In: F. Fournier, J. Mendling (eds.) Business Process Management Workshops - BPM 2014 International Workshops,Lecture Notes in Business Information Processing, vol.... on Pith
  20. [20] García-Valles, F., Colom, J.M.: Implicit places in net systems. In: Proceedings of the 8th International Workshop on Petri Nets and Per- formance Models, PNPM 1999, pp. 104–113. IEEE Computer Society (1999) on Pith
show 26 more
  1. [21] Groß, F.C., Mannel, L.L., van der Aalst, W.M.P.: Enhancing the appli- cability of the est-miner: Efficient precision-guided implicit place avoid- ance. In: 5th International Conference on Process Mining, ICPM 2023, pp. 121–128. IEEE (2023) on Pith
  2. [22] Leemans, S.J.J., Fahland, D., vanderAalst, W.M.P.: Discoveringblock- structured process models from event logs - A constructive approach. In: J.M. Colom, J. Desel (eds.) Application and Theory of Petri Nets and Concurrency - 34th International Conference, Petri NETS 2013, Lect... on Pith
  3. [23] Leemans, S.J.J., Fahland, D., vanderAalst, W.M.P.: Discoveringblock- structured process models from event logs containing infrequent be- haviour. In: M. La Rosa, P. Soffer (eds.) Business Process Manage- ment Workshops - BPM 2013 International Workshops,Lecture Notes in Busine... on Pith
  4. [24] Leemans, S.J.J., Fahland, D., van der Aalst, W.M.P.: Scalable process discovery and conformance checking. Softw. Syst. Model.17(2), 599– 631 (2018) on Pith
  5. [25] Leemans, S.J.J., Tax, N., ter Hofstede, A.H.M.: Indulpet miner: Com- bining discovery algorithms. In: H. Panetto, et al. (eds.) On the Move to Meaningful Internet Systems. OTM 2018 Conferences,Lecture Notes in Computer Science, vol. 11229, pp. 97–115. Springer (2018) 56 on Pith
  6. [26] de Leoni, M., Mannhardt, F.: Road traffic fine management process (2015) on Pith
  7. [27] Li, J., Liu, D., Yang, B.: Process mining: Extending alpha-algorithm to mine duplicate tasks in process logs. In: APWeb/WAIM Workshops (2007) on Pith
  8. [28] Lipton, R.J.: The reachability problem requires exponential space. Tech. Rep. 62, Department of Computer Science, Yale University (1976) on Pith
  9. [29] Mannel, L.L., van der Aalst, W.M.P.: Finding complex process- structures by exploiting the token-game. In: S. Donatelli, S. Haar (eds.) Application and Theory of Petri Nets and Concurrency - 40th International Conference, Petri NETS 2019,Lecture Notes in Computer Science, vol.... on Pith
  10. [30] Mannel, L.L., van der Aalst, W.M.P.: Finding uniwired Petri nets using est-miner. In: C.D. Francescomarino, R.M. Dijkman, U. Zdun (eds.) Business Process Management Workshops - BPM 2019 International Workshops,Lecture Notes in Business Information Processing, vol. 362, pp. 224... on Pith
  11. [31] Mannel, L.L., van der Aalst, W.M.P.: Discovering process models with long-term dependencies while providing guarantees and handling infre- quent behavior. In: L. Bernardinello, L. Petrucci (eds.) Application and Theory of Petri Nets and Concurrency - 43rd International Confere... on Pith
  12. [32] Mannel, L.L., Bergenthum, R., van der Aalst, W.M.P.: Removing im- plicit places using regions for process discovery. In: Proceedings of the International Workshop on Algorithms & Theories for the Analysis of Event Data 2020,CEUR Workshop Proceedings, vol. 2625, pp. 20–32. CEUR... on Pith
  13. [33] Mannel, L.L., Epstein, Y., van der Aalst, W.M.P.: Improving the state- space traversal of the est-miner by exploiting underlying log structures. In: A. del-Río-Ortega, H. Leopold, F.M. Santoro (eds.) Business Pro- cessManagementWorkshops-BPM2020InternationalWorkshops,Lec- ture... on Pith
  14. [34] Mannhardt, F.: Sepsis cases - event log (2016) 57 on Pith
  15. [35] Mannhardt, F., de Leoni, M., Reijers, H.A.: The multi-perspective pro- cess explorer. In: Proceedings of the BPM Demo Session 2015,CEUR Workshop Proceedings, vol. 1418. CEUR-WS.org (2015) on Pith
  16. [36] de Medeiros, A.K.A., Weijters, A.J.M.M., van der Aalst, W.M.P.: Ge- netic process mining: An experimental evaluation. Data Min. Knowl. Discov.14(2), 245–304 (2007) on Pith
  17. [37] Rozinat, A., van der Aalst, W.M.P.: Conformance checking of processes based on monitoring real behavior. Inf. Syst.33(1), 64–95 (2008) on Pith
  18. [38] Tiwari, A., Turner, C.J., Majeed, B.A.: A review of business process mining: State-of-the-art and future trends. Bus. Process. Manag. J. 14(1), 5–22 (2008) on Pith
  19. [39] Tour, A., Polyvyanyy, A., Kalenkova, A.A.: Agent system mining: Vi- sion, benefits, and challenges. IEEE Access9, 99480–99494 (2021) on Pith
  20. [40] Tsai, C., Jen, H., Chen, I.: Time-interval process model discovery and validation - A genetic process mining approach. Appl. Intell.33(1), 54–66 (2010) on Pith
  21. [41] Weijters, A.J.M.M., Ribeiro, J.T.S.: Flexible heuristics miner (FHM). In: 2011 IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2011, pp. 310–317. IEEE (2011) on Pith
  22. [42] Wen, L., vanderAalst, W.M.P., Wang, J., Sun, J.: Miningprocessmod- els with non-free-choice constructs. Data Min. Knowl. Discov.15(2), 145–180 (2007) on Pith
  23. [43] Wen, L., Wang, J., van der Aalst, W.M.P., Huang, B., Sun, J.: Mining process models with prime invisible tasks. Data Knowl. Eng.69(10), 999–1021 (2010) on Pith
  24. [44] vanderWerf, J.M.E.M., vanDongen, B.F., Hurkens, C.A.J., Serebrenik, A.: Process discovery using integer linear programming. In: K.M. van Hee, R. Valk (eds.) Applications and Theory of Petri Nets, 29th In- ternational Conference, Petri NETS 2008,Lecture Notes in Computer Scienc... on Pith
  25. [45] van Zelst, S.J., van Dongen, B.F., van der Aalst, W.M.P.: Avoiding over-fitting in ILP-based process discovery. In: H.R. Motahari-Nezhad, 58 J. Recker, M. Weidlich (eds.) Business Process Management - 13th In- ternational Conference, BPM 2015,Lecture Notes in Computer Science,... on Pith
  26. [46] van Zelst, S.J., van Dongen, B.F., van der Aalst, W.M.P., Verbeek, H.M.W.: Discovering workflow nets using integer linear programming. Computing100(5), 529–556 (2018) 59 on Pith
Cited by 1 Pith paper
How the literature uses it

What this paper does with the arXiv papers it cites

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

What it cites

Cites as background 1 paper

Challenges and discussion

Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.

No comments yet.

Sign in to comment