Pith. sign in

REVIEW 3 major objections 5 minor 29 references

A principled closure framework for higher-order SIS epidemic models on networks

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper establishes a network-dependent closure operator that closes the exact SIS hierarchy at the triplet level and derives three existing higher-order SIS models as special cases, exposing assumptions their original derivations left i

desk verdict A genuinely systematic closure operator that unifies three higher-order SIS models and exposes hidden assumptions in the inter-order model; the formal derivations hold, but the framework's predictive value is limited by an untested mean-field locality axiom and the absence of numerical validation. read the letter →

arxiv 2607.18003 v1 pith:MZPWYMUU submitted 2026-07-20 physics.soc-ph cond-mat.stat-mech

Kevin Teo , Peter L Simon , István Zoltán Kiss This is my paper
classification physics.soc-phcond-mat.stat-mech MSC 05C8234C6092D30
keywords higher-ordernetworksSISepidemicmomentclosureKirkwoodapproximationhyperedgemean-fieldmodelsoperatortriplet
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to turn heuristic moment closure for higher-order SIS epidemics on networks into a systematic, topology-aware procedure. It defines a closure operator that takes any four- or five-node motif, applies a hyperedge-tagging step and an edge-absence factorization step on top of the classical superposition approximation, and returns a product of single, pair, and triplet joint probabilities. This closes the exact microscopic rate equations at triplet level. The same operator, under specific structural and homogeneity assumptions, reproduces three existing mean-field higher-order SIS models; doing so reveals that the inter-order model's single overlap parameter is insufficient and must be replaced by a full distribution, a stronger sparsity condition, and a collapse of hyperedge isomorphism classes. A sympathetic reader would care because the framework converts a combinatorial wall of possible closures into a finite algorithm and makes previously invisible assumptions explicit.

What carries the argument

The closure map C_{S,E,H} = (∏_{e∉E} Φ_e) ∘ (∏_{h∈H} λ_h) composed with the (truncated) Kirkwood superposition approximation. λ_h ('hyperedge operator') converts a naive triangle-class triplet into its hyperedge-class counterpart, preserving a three-body correlation that never decomposes under edge removal. Φ_e ('independence operator') implements the paper's core factorization rule: if removing an edge disconnects a node from the rest of the motif, that node's state is independent of the others. The order matters — λ first, then Φ — and both families commute internally, so the result is unique and algorithmic.

What would settle it

Simulate exact SIS on a network containing a four-node motif {i,j,k,l} with edges {i,j},{j,k},{k,l} (a path) and no edge {i,k}, but where i and k both connect to a common high-degree infected hub outside the motif; measure ⟨XiXk⟩ and compare with the closure's factorized prediction ⟨Xi⟩⟨Xk⟩. If the correlation persists and the closed system's trajectories diverge from simulation, the factorization premise is falsified. Alternatively, for the inter-order model, construct two networks with the same α but different θ distributions and show the closure-operator system (tracking θ) and the original

Watch

Extended reading notes

Core claim

On the paper's own terms: the central discovery is that every higher-order SIS closure can be generated by one operator C = (∏ Φ_e) ∘ (∏ λ_h), where λ_h tags a hyperedge triplet so it survives as a correlated unit and Φ_e factorizes a joint probability when removing an edge disconnects the motif. Applied after the Kirkwood or truncated Kirkwood approximation, C rewrites every quadruplet and quintuplet term in the exact node, pair and triplet rate equations as products of lower-order probabilities. The paper proves that a maximal-clique model, a pair-based simplicial model, and an inter-order overlap model each emerge as special cases under explicit structural and mean-field assumptions. The

Load-bearing premise

The whole closure rests on equation (7): if removing an edge from a motif disconnects a node from the rest, then that node's state is independent of the others and the joint probability factorizes; this mean-field assertion can fail when correlations arrive through indirect routes that leave no direct edge in the motif.

Editorial extensions

If this is right

  • Closing the hierarchy at triplet level becomes a finite, automatic procedure for any four- or five-node motif, removing the combinatorial wall that previously forced manual closures.
  • Recovering existing models pins down their hidden assumptions: the inter-order model needs the full edge-count distribution θ, a stronger sparsity condition, and a collapse of hyperedge isomorphism classes; the maximal-clique model's population-level average hides a node-level homogeneity requirement.
  • The pair-based simplicial model is obtained exactly under regularity, the simplicial constraint, pair-level closure, and dynamical plus topological homogeneity — a complete, explicit ingredient list.
  • The framework identifies which mean-field assumptions are essential (topological homogeneity for aggregate models) and which can be relaxed, e.g., lifting the hyperedge-class collapse by tracking θ-resolved hyperedge triplet probabilities.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the factorization-on-disconnection premise is violated — for example, when two nodes with no direct edge are statistically correlated through a common infected neighbor or shared environmental source — the closed equations will understate those correlations; a targeted simulation comparison on such networks would reveal the boundary of the framework.
  • Because α is only the first moment of θ, networks with identical α but different θ distributions should produce measurably different epidemic dynamics in the triplet-resolved closure; the inter-order model's good simulation fit may reflect the specific θ of the tested hypergraph ensembles rather than the sufficiency of α.
  • The same operator recipe — enumerate motif isomorphism classes, write linear density relations using complement identities, promote partial sums to global averages — suggests a direct path to closures for larger hyperedges (size >3) by truncating the Kirkwood expansion more than two orders, and to other dynamics such as SIR or voter models.
  • The 'promoted sum' assumption (topological homogeneity) is unlikely to hold on strongly heterogeneous networks; one testable extension is a heterogeneous version that keeps degree/hyperedge-degree classes separate, which the framework's isomorphism-class bookkeeping can accommodate.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper develops a systematic moment-closure framework for SIS epidemics on networks with both pairwise edges and three-node hyperedges. Starting from exact microscopic rate equations for single nodes, pairs, and triplets (Eqs. (1)–(4)), the authors define a closure operator (Definition 3.12) that maps any four- or five-node joint probability to products of lower-order probabilities. The operator composes a hyperedge-tagging step with an edge-removal independence step, applied on top of (truncated) Kirkwood approximations. The main claims are: (i) the operator recovers three existing higher-order SIS models as special cases—Burgio et al.'s maximal-clique model, Malizia et al.'s pair-based model, and Malizia et al.'s inter-order model; and (ii) the inter-order recovery exposes hidden assumptions in the original derivation, namely a need for the overlap distribution θ, a stronger sparsity condition, and a hyperedge-class collapse. The paper is primarily a formal derivation exercise; it explicitly declines to provide computational tests.

Significance. If the formal recovery claims hold, this is a valuable contribution to the higher-order epidemic modeling literature. The closure operator provides a constructive, parameter-free procedure for generating closures from local topology, replacing hand-derived heuristics with a unified algebraic framework. The term-by-term derivations in the appendices are a particular strength, as they make the structural and dynamical assumptions behind each recovered model explicit. The most striking result is the analysis of the inter-order model, showing that the scalar overlap parameter α is insufficient without additional distributional and sparsity assumptions. This is a falsifiable and useful structural insight. The paper's main weakness is that the central locality premise—factorization upon disconnection, Eq. (7)—is not numerically or analytically validated, and the paper itself concedes the absence of computational tests. Thus the formal machinery is convincing, but its advertised utility as a generator of reliable new models remains open.

major comments (3)
  1. [§3.4, Eq. (7)] The independence operator Φ, and hence every closure produced by Eqs. (20)–(21), rests on Eq. (7): if removing an edge disconnects a node from the rest of the motif, the joint probability factorizes. This is a mean-field/locality assertion. In SIS on clustered or heterogeneous networks, correlations can be induced by common neighbors or indirect routes that are not represented by a single edge inside the motif, so the factorization can fail. The paper provides no numerical or analytical evidence delimiting when Eq. (7) holds; Section 7 explicitly states that no computational tests were undertaken. Since the framework is advertised as a 'principled method of generating new models', this is a load-bearing gap for the framework's predictive reliability. I ask for at least one benchmark—e.g., implement a class-resolved hyperedge-pendant or bowtie closure generated by the operator and compare
  2. [Proposition 3.11 / Definition 3.12] Definition 3.12 defines the closure map as a composition of independence operators ∏Φ_e, and its well-definedness depends on Proposition 3.11, which asserts idempotency and commutativity of the Φ operators. The proof of commutativity for overlapping pairs is asserted rather than demonstrated: it states that removing two edges in either order produces the same final configuration. For non-hyperedge triplets, however, the demotion chain in Definition 3.10 includes a decomposition step, and intermediate expressions contain pair and single factors that can be acted on by later operators. A rigorous proof should provide a full case analysis or a confluence argument for this rewriting system. As written, the commutativity claim is plausible but not fully established, and it underpins the uniqueness of the closure output.
  3. [§6.3, Assumptions 6.25/6.26 and Theorem 6.36] The paper claims that the inter-order model of Malizia et al. implicitly invokes additional structural assumptions beyond those stated, namely strong sparsity (Assumption 6.25) and hyperedge-class collapse (Assumption 6.26). What is shown is that the authors' particular derivation route from the microscopic equations requires these assumptions. That is a sufficient-condition argument, not a demonstration that no other derivation of Eqs. (57)–(60) could proceed under the weaker Assumption 6.16. Since the paper's central 'hidden assumptions' claim is stated in strong terms (Abstract, Section 7), the reader needs either a concrete network/parameter regime where Assumption 6.16 without 6.25 yields a different closed system or different predictions, or a rephrasing of the conclusion as 'our derivation requires' rather than 'the original derivation implicitly invokes'.
minor comments (5)
  1. [Lemma 6.14] The proof states that applying the truncated Kirkwood approximation \tildeκ to a four-node simplex-plus-external-node motif yields the displayed pair-only expressions. In fact, \tildeκ on four nodes produces triplet factors; the displayed formulas are obtained only after additionally applying the pair-level wedge/triangle closures (42). Please state this two-step reduction explicitly to avoid the impression that \tildeκ itself contains no triplet terms.
  2. [Eq. (26)] The displayed derivation with '. . .' and a down arrow is difficult to read. Rewrite as a clean two- or three-line derivation showing the replacement of node-indexed probabilities by class averages.
  3. [Assumption 6.25] The phrase 'any wedge shares at most one leaf with any triangle or hyperedge' is ambiguous. Define 'leaf' precisely (the two non-central nodes of a wedge) and clarify whether 'shares a leaf' means sharing a single node or sharing a node that plays a leaf role in both motifs.
  4. [Definition 3.12 / Eqs. (20)–(21)] The products of operators would benefit from an explicit composition-order convention (e.g., right-to-left). Also clarify that for absent hyperedges the factor \bar H_{xyz} is the identity operator, not a numerical prefactor acting on probabilities.
  5. [Eq. (33)] The term '−(1 + β(1))' in the BMC target equation: please clarify the normalization of rates so that the factor '1' is dimensionless; as written it appears to conflate rates and probabilities.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the closure framework is derived from explicitly stated assumptions, and model recovery is a consistency check rather than a fitted prediction.

full rationale

The paper's central construct is a closure operator C = (∏Φ_e)∘(∏λ_h) built on the Kirkwood approximation plus explicitly stated topological rules (Eq. (7), Definitions 3.8 and 3.10). These rules are assumptions, not outputs of the derivation; the paper does not fit parameters to data or rename an empirical pattern as a prediction. The recovery of Burgio et al., Malizia et al. pair-based, and Malizia et al. inter-order models is performed by deriving the target equations and closures from the microscopic rate equations (1)–(4), with structural constraints and homogeneity assumptions stated in each subsection. Even the inter-order case, which is the paper's critical finding, does not smuggle the target back in: the authors show that the explicit assumptions of [16] are insufficient, and that recovering the model requires additional θ-parameterization, strong sparsity, and hyperedge-class collapse — a conditional result, not a tautology. The cited target papers include a coauthor of the present paper, but they are used as objects of derivation, not as authority; no load-bearing claim rests on a self-citation or uniqueness theorem. Eq. (7) is a mean-field locality premise whose empirical validity is not tested, but as a stated assumption it does not make the derivation circular. The paper itself flags the absence of computational tests and the Kirkwood basis, which is a limitation, not circularity.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

The framework itself introduces no fitted parameters or new physical entities. The structural descriptors (k1, k2, ϕ, δ, α, θ) are inputs inherited from the models being recovered. Assumptions 1-3 and 6-7 are standard in network mean-field theory; assumptions 4, 5 and 8 are new rules introduced by this paper, of which 8 is the most fragile because it is only justified by the need to match the IO model.

assumptions (8)
  • domain assumption SIS dynamics as a continuous-time Markov chain with independent Poisson processes (infection rates β1, β2 and recovery γ)
    Section 3.1: the microscopic rate equations (1)-(4) are marginal probabilities of the Kolmogorov forward equation. Standard for network epidemic models.
  • domain assumption Moment hierarchy truncated at triplets; quadruplet and quintuplet probabilities are closed via κ and κ̃
    Section 3.2: 'we truncate the hierarchy at the level of triplets'. This is the modeling choice that the closure operator implements, not a proven reduction.
  • domain assumption Kirkwood superposition approximation κ (Definition 3.4) and its truncated version κ̃ (Definition 3.6) provide the base ansatz for joint probabilities
    Borrowed from statistical mechanics [25,26]. The paper states in Section 7 that regimes where Kirkwood is a poor fit will affect the framework.
  • ad hoc to paper Statistical dependence is mediated only by transmission paths (Eq. 7): disconnecting a node from a motif implies factorization
    Section 3.4, first observation. This is the load-bearing premise behind the independence operator Φ and hence every closure the framework produces.
  • ad hoc to paper Hyperedges preserve three-node correlation regardless of pairwise edge removal; hyperedge triplets demote but never decompose (Definition 3.10)
    Section 3.4.3, 'Action on hyperedge triplets'. Central design rule for the hyperedge operator; not derived from the microscopic dynamics.
  • domain assumption Dynamical homogeneity (Assumption 5.1): joint state probabilities depend only on isomorphism class, not node labels
    Section 5.1. Standard mean-field assumption used to reduce label-specific equations.
  • domain assumption Topological homogeneity (Assumption 5.2): local subgraph counts equal network-wide averages (promoted sums)
    Section 5.2. Needed for the mean-field PBS and IO recoveries; exact only for regular networks.
  • ad hoc to paper Strong sparsity (Assumption 6.25) and hyperedge-class collapse (Assumption 6.26)
    Section 6.3.3. Introduced to close the IO derivation; the paper argues they are implicit in [16] but not stated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A principled closure framework for higher-order SIS epidemic models on networks." pith.science (2026). https://pith.science/paper/MZPWYMUU

@misc{pith2026260718003,
  author       = {Pith},
  title        = {Pith review of: A principled closure framework for higher-order SIS epidemic models on networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MZPWYMUU}},
  note         = {Machine review of arXiv:2607.18003}
}
read the original abstract

Susceptible-infected-susceptible (SIS) epidemic models on networks are governed by hierarchical moment equations where the dynamics of smaller subsystems depend on the state of larger ones. Moment closure approximations, which truncate this hierarchy by expressing higher-order state probabilities in terms of lower-order ones, are essential for obtaining tractable reduced systems. Higher-order networks, which extend the pairwise structure to include group interactions, introduce a combinatorial explosion of closure configurations, making systematic derivation harder. Consequently, existing higher-order SIS models are derived heuristically, where structural and dynamical assumptions underpinning their closures are not always apparent from the formulation alone. We develop a bottom-up derivation of higher-order SIS dynamics, building systematically from node-level equations to pairs, triplets, and three-body interactions. Central to our approach is a network-dependent closure operator that generates topologically appropriate approximations from local pairwise and triadic structure. Using this framework, we recover three existing higher-order SIS models--Burgio et al.'s maximal clique, Malizia et al.'s pair-based and inter-order models--as special cases, each arising under specific topological and dynamical assumptions. Our derivation reveals assumptions that are invisible from heuristic approaches: for instance, Malizia et al.'s inter-order overlap parameter is insufficient alone to express the model within our framework despite performing well against simulations, with the original derivation implicitly invoking additional structural assumptions. Our framework offers both a foundation for higher-order epidemic modeling and a constructive pathway for understanding the assumptions implicit in heuristically derived mean-field closures and provides a principled method of generating new models.

Figures

Figures reproduced from arXiv: 2607.18003 by the authors.

Figure 1
Figure 1. Diagrams of all possible 3-node sub-graphs with pairwise edges and hyperedges. Hyperedges are shown in blue in panels (a)–(d). 9 [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Example of 4 nodes with a hyperedge {i, j, k} and edges {i, j}, {i, k} and {k, l}. 4 Canonical closure approximations We derive several canonical closure approximations commonly used in the existing literature using our closure operator framework. The key insight is that closure expressions in the literature are often limited to well-known or simple motifs, but our closure operator framework permits a closure of any… view at source ↗
Figure 3
Figure 3. Schematic illustrating the pipeline used in deriving specific higher-order contagion models [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Allowed motif configurations under the explicit (weaker) assumptions of Assumption [PITH_FULL_IMAGE:figures/full_fig_p036_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

29 extracted references · 2 linked inside Pith

  1. [1]

    Oxford university press, 1991

    Roy M Anderson and Robert M May.Infectious diseases of humans: dynamics and control. Oxford university press, 1991

    1991

  2. [2]

    Spread of epidemic disease on networks.Physical review E, 66(1):016128, 2002

    Mark EJ Newman. Spread of epidemic disease on networks.Physical review E, 66(1):016128, 2002

    2002

  3. [3]

    Epidemic processes in complex networks.Reviews of modern physics, 87(3):925–979, 2015

    Romualdo Pastor-Satorras, Claudio Castellano, Piet Van Mieghem, and Alessandro Vespignani. Epidemic processes in complex networks.Reviews of modern physics, 87(3):925–979, 2015

    2015

  4. [4]

    Mathematics of epidemics on networks

    István Z Kiss, Joel C Miller, Péter L Simon, et al. Mathematics of epidemics on networks. Cham: Springer, 598(2017):31, 2017

    2017

  5. [5]

    Higher-order network analysis takes off, fueled by old ideas and new data.SIAM News Blog, 21(1), 2021

    Austin R Benson, David F Gleich, and Desmond J Higham. Higher-order network analysis takes off, fueled by old ideas and new data.SIAM News Blog, 21(1), 2021

    2021

  6. [6]

    What are higher-order networks?SIAM review, 65(3):686–731, 2023

    Christian Bick, Elizabeth Gross, Heather A Harrington, and Michael T Schaub. What are higher-order networks?SIAM review, 65(3):686–731, 2023

    2023

  7. [7]

    Networks beyond pairwise interactions: Structure and dynamics.Physics reports, 874:1–92, 2020

    FedericoBattiston, GiuliaCencetti, IacopoIacopini, VitoLatora, MaximeLucas, AlicePatania, Jean-Gabriel Young, and Giovanni Petri. Networks beyond pairwise interactions: Structure and dynamics.Physics reports, 874:1–92, 2020

    2020

  8. [8]

    Defining and classifying models of groups: The social ontology of higher-order networks.arXiv preprint arXiv:2507.02758, 2025

    Jonathan St-Onge, Randall Harp, Giulio Burgio, Timothy M Waring, Juniper Lovato, and Laurent Hébert-Dufresne. Defining and classifying models of groups: The social ontology of higher-order networks.arXiv preprint arXiv:2507.02758, 2025

    arXiv 2025

Show all 29 references
  1. [9]

    Dynamics on higher-order networks: A review.Journal of the Royal Society Interface, 19(188):20220043, 2022

    Soumen Majhi, Matjaž Perc, and Dibakar Ghosh. Dynamics on higher-order networks: A review.Journal of the Royal Society Interface, 19(188):20220043, 2022

    2022

  2. [10]

    Simplicial models of social contagion.Nature communications, 10(1):2485, 2019

    Iacopo Iacopini, Giovanni Petri, Alain Barrat, and Vito Latora. Simplicial models of social contagion.Nature communications, 10(1):2485, 2019

    2019

  3. [11]

    Social contagion models on hypergraphs.Physical Review Research, 2(2):023032, 2020

    Guilherme Ferraz de Arruda, Giovanni Petri, and Yamir Moreno. Social contagion models on hypergraphs.Physical Review Research, 2(2):023032, 2020. 50

    2020

  4. [12]

    Disentangling the role of heterogeneity and hyperedge overlap in explosive contagion on higher-order networks.Physical Review Letters, 135(20):207401, 2025

    Federico Malizia, Andrés Guzmán, Iacopo Iacopini, and István Z Kiss. Disentangling the role of heterogeneity and hyperedge overlap in explosive contagion on higher-order networks.Physical Review Letters, 135(20):207401, 2025

    2025

  5. [13]

    A pair-based approximation for simplicial contagion.Chaos, Solitons & Fractals, 199:116776, 2025

    Federico Malizia, Luca Gallo, Mattia Frasca, István Z Kiss, Vito Latora, and Giovanni Russo. A pair-based approximation for simplicial contagion.Chaos, Solitons & Fractals, 199:116776, 2025

    2025

  6. [14]

    Hyperedge overlap drives explosive transitions in systems with higher-order interactions.Nature communications, 16(1):555, 2025

    Federico Malizia, Santiago Lamata-Otín, Mattia Frasca, Vito Latora, and Jesús Gómez- Gardeñes. Hyperedge overlap drives explosive transitions in systems with higher-order interactions.Nature communications, 16(1):555, 2025

    2025

  7. [15]

    Triadic approximation reveals the role of interaction overlap on the spread of complex contagions on higher-order networks.Physical Review Letters, 132(7):077401, 2024

    Giulio Burgio, Sergio Gómez, and Alex Arenas. Triadic approximation reveals the role of interaction overlap on the spread of complex contagions on higher-order networks.Physical Review Letters, 132(7):077401, 2024

    2024

  8. [16]

    Nested hyperedges promote the onset of collective transitions but suppress explosive behavior.arXiv preprint arXiv:2601.10522, 2026

    Federico Malizia, Andrés Guzmán, Federico Battiston, and István Z Kiss. Nested hyperedges promote the onset of collective transitions but suppress explosive behavior.arXiv preprint arXiv:2601.10522, 2026

    2026 arXiv

  9. [17]

    Deterministic epidemiological models at the individual level.Journal of Mathematical Biology, 57(3):311–331, 2008

    Kieran J Sharkey. Deterministic epidemiological models at the individual level.Journal of Mathematical Biology, 57(3):311–331, 2008

    2008

  10. [18]

    Deterministic epidemic models on contact networks: Correlations and unbiological terms.Theoretical population biology, 79(4):115–129, 2011

    Kieran J Sharkey. Deterministic epidemic models on contact networks: Correlations and unbiological terms.Theoretical population biology, 79(4):115–129, 2011

    2011

  11. [19]

    Springer International Publishing, Cham, 2016

    Christian Kuehn.Moment Closure—A Brief Review, pages 253–271. Springer International Publishing, Cham, 2016

    2016

  12. [20]

    Correlation equations and pair approximations for spatial ecologies.Advanced ecological theory: principles and applications, pages 100–142, 1999

    DA Rand. Correlation equations and pair approximations for spatial ecologies.Advanced ecological theory: principles and applications, pages 100–142, 1999

    1999

  13. [21]

    The effects of local spatial structure on epidemiological invasions

    Matthew J Keeling. The effects of local spatial structure on epidemiological invasions. Proceedings of the Royal Society of London. Series B: Biological Sciences, 266(1421):859–867, 1999

    1999

  14. [22]

    A motif-based approach to network epidemics.Bulletin of Mathematical Biology, 71(7):1693–1706, 2009

    Thomas House, Geoffrey Davies, Leon Danon, and Matt J Keeling. A motif-based approach to network epidemics.Bulletin of Mathematical Biology, 71(7):1693–1706, 2009

    2009

  15. [23]

    Mean-field models of dynamics on networks via moment closure: An automated procedure.Physical Review E, 106(5):054312, 2022

    Bert Wuyts and Jan Sieber. Mean-field models of dynamics on networks via moment closure: An automated procedure.Physical Review E, 106(5):054312, 2022

    2022

  16. [24]

    Enumeration of hypergraphs.Discrete Mathematics, 45(2- 3):239–243, 1983

    Malti Hedge and MR Sridharan. Enumeration of hypergraphs.Discrete Mathematics, 45(2- 3):239–243, 1983

    1983

  17. [25]

    Statistical mechanics of fluid mixtures.The Journal of chemical physics, 3(5):300–313, 1935

    John G Kirkwood. Statistical mechanics of fluid mixtures.The Journal of chemical physics, 3(5):300–313, 1935

    1935

  18. [26]

    Physical nature of higher-order mutual information: Intrinsic correlations and frustration.Physical review E, 62(3):3096, 2000

    Hiroyuki Matsuda. Physical nature of higher-order mutual information: Intrinsic correlations and frustration.Physical review E, 62(3):3096, 2000

    2000

  19. [27]

    The mathematics of infectious diseases.SIAM review, 42(4):599–653, 2000

    Herbert W Hethcote. The mathematics of infectious diseases.SIAM review, 42(4):599–653, 2000. 51

    2000

  20. [28]

    Springer, 5th edition, 2017

    Reinhard Diestel.Graph Theory, volume 173 ofGraduate Texts in Mathematics. Springer, 5th edition, 2017

    2017

  21. [29]

    Elsevier, 2014

    Judea Pearl.Probabilistic reasoning in intelligent systems: networks of plausible inference. Elsevier, 2014. 52

    2014

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.