Pith. sign in

REVIEW 3 major objections 5 minor 48 references

Hierarchical Clustering of Networks via Hierarchical Distance Matrices

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read This paper proves that when the population-level distance matrix of a collection of networks satisfies a recursive separation condition, a spectral top-down clustering algorithm exactly recovers the latent binary tree of clusters, and a two

desk verdict New HDM framework and NHC-TST algorithm with sound central theory; the stress-test note's counterexample is wrong, and remaining concerns are patchable. Deserves peer review. read the letter →

arxiv 2607.29296 v1 pith:544LXJYT submitted 2026-07-31 stat.ME

Li Chen , Nathaniel Josephs , Eric D. Kolaczyk , Lizhen Lin This is my paper · ORCID
classification stat.ME MSC 62H30
keywords HierarchicalDistanceMatrixnetworkclusteringspectraltwo-sampletestlatenthierarchygraphonestimationmultiplenetworkstop-down
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 authors introduce the Hierarchical Distance Matrix, a structural condition on the population-level pairwise distances between networks: at every internal node of an unknown binary tree, every between-branch distance must exceed every within-branch distance. They show that under this condition, the sign pattern of the dominant eigenvector of the associated Laplacian encodes the correct bipartition, so recursively splitting via spectral clustering exactly reconstructs the full hierarchy. In the empirical setting where distances are estimated from link probabilities, they prove sign recovery with high probability and, when a graph-based two-sample test is used as a stopping rule, exact recovery of the hierarchy with probability tending to one. A reader should care because this gives a formal, provably consistent way to recover latent hierarchical organization among populations of networks, without the user specifying the number of clusters or tree depth in advance.

What carries the argument

The central object is the Hierarchical Distance Matrix (HDM): a symmetric population-level distance matrix that, after a suitable permutation, can be recursively block-partitioned so that the off-diagonal (between-branch) block strictly dominates the diagonal (within-branch) blocks at every internal node. The carrying mechanism is the spectral sign check: for the Laplacian L_D = diag(D1)-D, the dominant eigenvector has entries whose signs exactly separate the two branches; the proof uses a positive-negative block matrix lemma showing that this eigenvector has a strict sign pattern, plus Davis-Kahan bounds to transfer the result to estimated distances.

What would settle it

Generate a population of networks from an unbalanced tree where a deep split has a larger population distance than a shallow split (violating HDM monotonicity), e.g., the mirrored version of Section 4.3, and check whether the dominant Laplacian eigenvector's signs separate the true root branches; the paper's own Theorem 1 would not apply, and the split should fail at the root.

Watch

Extended reading notes

Core claim

The paper's central claim is that a population of networks generated from K latent link-probability matrices arranged on an unknown binary tree is exactly recoverable — both cluster labels and the tree topology — whenever the population-level distance matrix D is a Hierarchical Distance Matrix. This means that after permuting rows and columns, every internal submatrix can be split into two diagonal blocks and an off-diagonal block such that the off-diagonal entries are strictly larger than all entries inside the diagonal blocks, recursively down to the leaves. Under that condition the sign pattern of the dominant eigenvector of the Laplacian L_D = diag(D1)-D coincides with the first bipartit

Load-bearing premise

The load-bearing premise is that at every internal node of the latent tree, every distance between networks on different branches is strictly larger than every distance within either branch (the HDM monotonicity condition); if this recursive separation fails at any node, the spectral sign pattern need not align with the true bipartition and the exact-recovery guarantees collapse.

Editorial extensions

If this is right

  • Theoretical exact recovery: if D is an HDM, NHCL recovers every bipartition exactly (Theorem 1), and recursively the full tree (Corollary 1); NHCD requires the additional separation conditions in Theorem 2/Corollary 2.
  • Statistical consistency: with a uniformly consistent estimator of link probabilities (e.g., neighborhood smoothing), the empirical eigenvector's sign pattern matches the population one with probability at least (1-n^{-ε})^m, making NHCL asymptotically exact (Theorem 3).
  • Fully adaptive hierarchy: NHC-TST requires no prior knowledge of the number of clusters or tree depth; the two-sample test decides where to stop, and Theorem 4 shows the full hierarchy is recovered with probability at least 1 - Σ(δ_x=1)(m_x n^{-ε} + β_x) - 2L α_n, which tends to 1 under asymptotic power one.
  • Empirical superiority: in simulations with hierarchical smooth graphons and hierarchical SBMs, NHC-TST variants achieve near-perfect cluster membership (NMI) and topology recovery (adapted CPCC) across dense and sparse regimes, outperforming existing flat and hierarchical network-clustering baselines.
  • Interpretable multi-resolution structure: on the global migration dataset, the method recovers a temporal hierarchy that isolates the COVID-19 crisis and distinguishes post-COVID recovery from the pre-pandemic regime, structure invisible to flat clustering.

Reading between the lines

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

  • A direct corollary of the proof structure is that any consistent estimator of the population distance matrix — not only Frobenius distances of link probabilities — can be plugged into the same spectral splitting scheme; the framework is tied to the distance matrix, not to graphon estimation per se.
  • The HDM condition resembles a 'tree metric with layer-separated ranges' and is likely satisfied when signal components decay geometrically down a tree, suggesting the method transfers to hierarchical stochastic block models, topic hierarchies, and other nested-network generative models beyond the two simulation families.
  • The finite-sample bottleneck is the two-sample test's power at internal nodes; since the stopping rule must reject at each true internal split, the effective sample size per node limits recoverable depth — a user-facing trade-off the paper does not fully explore.
  • Section 4.3's mild HDM violation still yields near-perfect CPCC, hinting that the strict monotonicity condition may be stronger than necessary; an explicit tolerance theorem for 'approximately HDM' matrices would be a natural next step.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 introduces the Hierarchical Distance Matrix (HDM), a recursive separation condition on population-level distance matrices that encodes a latent binary tree over populations of networks. It proposes NHC-TST, a top-down algorithm that recursively partitions networks by spectral clustering and uses a graph-based two-sample test as a data-driven stopping rule. The main theoretical claims are: (i) Theorem 1/Corollary 1, exact recovery of each population-level bipartition and hence the full hierarchy by the Laplacian spectral method NHCL when the true distance matrix is an HDM; (ii) Theorem 2/Corollary 2, a parallel guarantee for the distance-matrix method NHCD; (iii) Theorem 3, consistency of NHCL with an estimated distance matrix under uniform link-probability estimation error; and (iv) Theorem 4, exact recovery with probability tending to 1 for the full NHCL-TST procedure. The paper also reports extensive simulations and an application to global migration networks.

Significance. If the theoretical claims are established, the HDM framework is a genuinely useful formalization of latent hierarchy for network populations, and the proposed recursive, test-based procedure is an appealing alternative to flat clustering methods. The paper's strengths include a clearly stated structural model, explicit separation assumptions, a data-driven stopping rule that avoids prespecifying the number of clusters, and a broad simulation study with a hierarchy-aware cophenetic evaluation metric. The migration application is also interpretable. The central proofs, however, currently contain load-bearing gaps: the proof of Theorem 1 does not rigorously transfer the sign pattern from an auxiliary matrix to the Laplacian, and the proof of Theorem 2 contains an invalid inequality in the Rayleigh-quotient argument. These gaps appear repairable with existing ingredients, so the paper is promising but not yet ready in its present form.

major comments (3)
  1. [Appendix S.2 (proof of Theorem 1)] The proof constructs B = κ11^T + L_D and invokes Lemma 1 to give the dominant eigenvector u1 a mixed sign pattern. To conclude that u1 is the dominant eigenvector of L_D, the proof must show that u1 lies in the subspace orthogonal to 1. The text argues: 'Observe that the signs of 1 are the same, implying u1 ≠ 1. Consequently, u1 must belong to the set of remaining eigenvectors orthogonal to 1.' The inference from u1 ≠ 1 to u1 ⊥ 1 requires that the eigenvalue κm of B is distinct from λ1(B) and that the dominant eigenspace is simple; this is not established. The gap is repairable: since κ ∈ (β0, α1), the matrix SBS with S = diag(I_{m0}, -I_{m1}) is entrywise positive, so Perron-Frobenius gives a simple dominant eigenvector; alternatively Lemma 5 gives λ1(L_D) ≥ mα1 > κm. But as written, Theorem 1, Corollary 1, and the population-level basis of Theorem 4 are not proven.
  2. [Appendix S.3 (proof of Theorem 2)] The central displayed inequality in the mixed-sign case and its predecessor replace sums of products |u_i||u_j| by cardinalities of sign classes. For the positive within-block terms, the upper bound by β0 times the product of counts is valid because each |u_i| ≤ 1. But for the between-block terms, the proof effectively lower-bounds Σ_{i∈I+,j∈J+} |u_i||u_j| by m_{0+} m_{1+} (and similarly for the other cross term), which is not valid for a unit vector; such a sum can be arbitrarily small. Consequently the derived condition α1 > β0 + M is not established by the displayed algebra. Since Theorem 2 is the exact-recovery guarantee for the NHCD variant and underpins Corollary 2, this needs a rigorous argument or a corrected sufficient condition.
  3. [Appendix S.4 (Lemma 5 and Definition 4)] Lemma 5, used repeatedly in the proof of Theorem 3 and in any repair of Theorem 1, asserts after constructing the HCMs Hα and Hβ that Hα,ij ≤ Dij ≤ Hβ,ij for all pairs. This entrywise comparison is not immediate from Definition 4, which specifies only the constant between-block scalars at each internal node and does not explicitly define the entries of the diagonal blocks for pairs that split at lower levels. The comparison can likely be made correct by defining Hβ,ij as the maximal between-block constant on the path where i,j first separate and Hα,ij analogously, but the proof as written omits this construction. Because the lemma is load-bearing, it should be stated and proved explicitly.
minor comments (5)
  1. [Algorithm 1, lines 7–8] The pseudocode is ambiguous: line 7 tests a p-value but then line 8 calls BIPARTITION again although a partition was already produced in line 3. Please clarify whether BIPARTITION simply returns the existing split or refers to a different procedure, and make the stopping logic explicit.
  2. [Lemma 6, Eq. (S.9)] The displayed bounds contain a typographical ambiguity: 'c1√mc2 ≤ |u_i| ≤ c2/2√m' should presumably read c1/(c2√m) ≤ |u_i| ≤ c2/(2√m).
  3. [Proof of Theorem 4, final line] The notation 'lim_{n→∞}^{m_x=o(n^ε)} P(E_T)=1' is malformed; E_T is not defined. Also, internal nodes with δ_x = 1 cannot occur at depth L, so the summation over l = 0,...,L should be adjusted or explained.
  4. [Assumption 7] Assumption 7 (λ1(L_D) > deg_i(D) for all i) is used in the proof of Lemma 6 and Theorem 3, but no discussion is given of when it is satisfied. A brief comment that it is a mild regularity condition, or a sufficient condition, would improve the paper.
  5. [References] The reference list contains formatting inconsistencies, e.g. 'A. Spielman, D. (2009). Lecture Notes on Random Matrix Theory.'

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; central recovery results are derived from stated HDM assumptions, with only minor reliance on the authors' own prior two-sample test.

full rationale

The derivation chain is essentially non-circular. The Hierarchical Distance Matrix condition (Definitions 1-2) is a structural assumption on the population-level distance matrix: it requires nested within/between block inequalities at every non-terminal prefix, but it does not assert the spectral sign-separation conclusion. Theorem 1 and Corollary 1 derive the bipartition recovery from this assumption; the assumption and the conclusion are not the same object. Similarly, Theorem 3 is deliberately modular: the paper states that it 'does not rely on a specific construction of the link probability matrix estimators. Instead, it applies to any estimators satisfying the uniform error bound in Assumption 4.' This is the opposite of fitting a parameter and then calling it a prediction. Theorem 4 is stated conditionally on the embedded two-sample test being asymptotically powerful, and the stopping rule is a plug-in component. The paper's reliance on Chen et al. (2024) for the MNBS two-sample test and on Zhao et al. (2019) for NBS/MNBS estimation rates is self-citation, but it is a modular citation to prior published methodology rather than a circular reduction: the cited results are not used to define the target hierarchy, and no fitted value is renamed as a prediction. The adapted CPCC is an evaluation metric, not part of the recovery derivation. The proof of Theorem 1 as written contains a serious gap: since B1 = kappa*m*1, the all-ones vector is an eigenvector of B, so the appeal to Lemma 1 for the dominant eigenvector does not obviously go through. However, that is a correctness/validity concern, not circularity: the HDM assumption does not assume the eigenvector sign pattern. Likewise, Section 4.3's 'mild departure from the HDM monotonicity' is an explicit caveat about assumption violation, not a circular step. Score 2 reflects the minor self-citation burden in the stopping-rule justification, while the central derivation remains self-contained relative to its stated assumptions.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The theoretical engine rests on the HDM separation assumption plus external estimators and tests. No new physical entities, forces, or particles are introduced. The most paper-specific assumptions are Assumptions 5-7, which are ad hoc technical conditions needed for the consistency proof, and the HDM monotonicity condition, which is the main domain assumption.

free parameters (3)
  • tau (binarization quantile) = 0.7 (chosen after scanning 0.2-0.8)
    The migration application binarizes directed weighted edges by retaining flows above the tau-th quantile; tau is a data-dependent threshold and the paper reports tau=0.7 as the representative case after examining 13 levels.
  • alpha_n (nominal significance level) = 1/n
    The stopping-rule significance level is set to 1/n in the simulations; it controls when the algorithm stops splitting and is a user-chosen tuning parameter rather than a derived quantity.
  • NBS/MNBS bandwidth and tuning parameters
    The neighborhood smoothing and modified neighborhood smoothing estimators used in Algorithm 1 require implementation choices; no concrete tuning values are given, which affects exact replication.
assumptions (6)
  • domain assumption The population distance matrix is an HDM: at every non-terminal node, all between-branch distances strictly exceed all within-branch distances (Definition 1 + Assumptions 1-2).
    This is the structural premise on which the exact recovery theorems rest. The paper itself notes mild violations in Section 4.3.
  • domain assumption Each observed network is an undirected simple graph with edges generated independently as Bernoulli given the cluster-specific link-probability matrix (Section 2.1).
    The probabilistic model under which the distance matrix and estimators are defined.
  • domain assumption NBS/MNBS estimators satisfy a uniform Frobenius error bound O(sqrt(log n/n)) (Zhang et al. 2017; Zhao et al. 2019).
    Assumption 4 and Remark 3 import this rate from cited prior work to control perturbation of the estimated distance matrix.
  • domain assumption The two-sample test of Chen et al. (2024) has asymptotic power one and controlled Type I error under the conditions used.
    Theorem 4's stopping-rule consistency is inherited directly from this test; the paper does not re-prove its power properties.
  • ad hoc to paper Technical Assumptions 5-7: Delta = omega(zeta), beta_1 = O(Delta), and lambda_1(L_D) > deg_i(D) for all i.
    These conditions are introduced specifically to make the proofs in Theorem 3 and Lemma 6 go through; they are not derived from data or external results and are not verified in simulations.
  • domain assumption The latent hierarchy is a binary tree with finite depth L.
    The model, algorithm, and theory use recursive bipartitions; the paper lists extension to non-binary trees as future work.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hierarchical Clustering of Networks via Hierarchical Distance Matrices." pith.science (2026). https://pith.science/paper/544LXJYT

@misc{pith2026260729296,
  author       = {Pith},
  title        = {Pith review of: Hierarchical Clustering of Networks via Hierarchical Distance Matrices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/544LXJYT}},
  note         = {Machine review of arXiv:2607.29296}
}
read the original abstract

Clustering populations of networks while recovering their latent hierarchical organization is a fundamental yet largely unexplored problem in network analysis. To formalize this, we introduce the Hierarchical Distance Matrix, a specific class of population-level distance matrices that encodes latent hierarchical organization through recursively nested distance separation, accommodating unbalanced tree depths. Building on this framework, we propose a fully data-driven top-down procedure: network hierarchical clustering based on two-sample testing (NHC-TST). The algorithm recursively splits networks via spectral clustering and uses a graph-based two-sample stopping rule. The procedure adaptively determines the branching structure without requiring prior knowledge of the number of clusters or tree depth. Theoretically, we establish exact recovery of the population-level hierarchical structure and statistical consistency in the empirical procedure. Simulation studies demonstrate highly accurate recovery of both cluster memberships and hierarchical relationships across a wide range of settings. Applied to a global migration dataset, NHC-TST uncovers interpretable multi-resolution temporal structures that are not revealed by conventional flat clustering approaches.

Figures

Figures reproduced from arXiv: 2607.29296 by the authors.

Figure 1
Figure 1. Hierarchical tree of the underlying link-probability matrices from which populations of networks are drawn. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Hierarchical monotonicity heatmaps (dense regime, [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Behavior comparison when the underlying networks are generated from hierarchical smooth graphons. Panels [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Behavior comparison when the underlying networks are generated from SBMs. Panels report (a) NMI; (b) the [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Behavior comparison when the underlying networks are generated from non-uniform hierarchical smooth [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Cluster assignments on migration networks. Top: Policy stringency. Bottom: Assignments. [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Pairwise interaction networks from NHCD-TST based on anomalous interaction intensities, highlighting the top 20 anomalous edges and associated countries. Qatar (QA), and the United Arab Emirates (AE)). Within this stable branch, cluster 00 specifically highlights stead…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 4 linked inside Pith

  1. [1]

    Graph evolution:

    Leskovec, Jure and Kleinberg, Jon and Faloutsos, Christos , journal=. Graph evolution:. 2007 , publisher=

    2007

  2. [2]

    Migration networks as a response to financial constraints:

    Stark, Oded and Jakubek, Marcin , journal=. Migration networks as a response to financial constraints:. 2013 , publisher=

    2013

  3. [3]

    Physical Review E—Statistical, Nonlinear, and Soft Matter Physics , volume=

    Extraction and analysis of traffic and topologies of transportation networks , author=. Physical Review E—Statistical, Nonlinear, and Soft Matter Physics , volume=. 2006 , publisher=

    2006

  4. [4]

    Handbook of Graphs and Networks: From Genome to Interne , pages=

    Complex networks in genomics and proteomics , author=. Handbook of Graphs and Networks: From Genome to Interne , pages=. 2002 , publisher=

    2002

  5. [5]

    Proceedings of the National Academy of Sciences of the United States of America , volume=

    Inferring friendship network structure by using mobile phone data , author=. Proceedings of the National Academy of Sciences of the United States of America , volume=. 2009 , publisher=

    2009

  6. [6]

    Journal of Neuroscience , volume=

    Hierarchical organization of human cortical networks in health and schizophrenia , author=. Journal of Neuroscience , volume=. 2008 , publisher=

    2008

  7. [7]

    Inferring international and internal migration patterns from

    Zagheni, Emilio and Garimella, Venkata Rama Kiran and Weber, Ingmar and State, Bogdan , booktitle=. Inferring international and internal migration patterns from

  8. [8]

    Proceedings of the National Academy of Sciences , volume=

    Quantifying seasonal population fluxes driving rubella transmission dynamics using mobile phone data , author=. Proceedings of the National Academy of Sciences , volume=. 2015 , publisher=

    2015

Show all 48 references
  1. [9]

    Unveiling hidden migration and mobility patterns in climate stressed regions:

    Lu, Xin and Wrathall, David J and Sunds. Unveiling hidden migration and mobility patterns in climate stressed regions:. Global Environmental Change , volume=. 2016 , publisher=

    2016

  2. [10]

    Geo-located

    Hawelka, Bartosz and Sitko, Izabela and Beinat, Euro and Sobolevsky, Stanislav and Kazakopoulos, Pavlos and Ratti, Carlo , journal=. Geo-located. 2014 , publisher=

    2014

  3. [11]

    Leveraging

    Zagheni, Emilio and Weber, Ingmar and Gummadi, Krishna , journal=. Leveraging. 2017 , publisher=

    2017

  4. [12]

    Human migration:

    S. Human migration:. International Journal of Data Science and Analytics , volume=. 2021 , publisher=

    2021

  5. [13]

    2023 , howpublished=

    2023

  6. [14]

    International Migration Flows to and from Selected Countries , author=

  7. [15]

    2016 , doi =

    Getting Skills Right: Assessing and Anticipating Changing Skill Needs , author=. 2016 , doi =

    2016

  8. [16]

    Taxon , volume=

    The comparison of dendrograms by objective methods , author=. Taxon , volume=. 1962 , publisher=

    1962

  9. [17]

    Proceedings of the 31st Advances in Neural Information Processing Systems , volume=

    On clustering network-valued data , author=. Proceedings of the 31st Advances in Neural Information Processing Systems , volume=

  10. [18]

    arXiv preprint arXiv:2307.09210 , year=

    Nested stochastic block model for simultaneously clustering networks and nodes , author=. arXiv preprint arXiv:2307.09210 , year=

    arXiv

  11. [19]

    Fan, Xing and Pensky, Marianna and Yu, Feng and Zhang, Teng , journal=

  12. [20]

    Proceedings of the National Academy of Sciences , volume=

    Measuring global migration flows using online data , author=. Proceedings of the National Academy of Sciences , volume=. 2025 , publisher=

    2025

  13. [21]

    Molecular Psychiatry , volume=

    The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism , author=. Molecular Psychiatry , volume=. 2014 , publisher=

    2014

  14. [22]

    Journal of machine learning research , volume=

    Cluster ensembles---a knowledge reuse framework for combining multiple partitions , author=. Journal of machine learning research , volume=

  15. [23]

    Journal of Classification , volume=

    Comparing partitions , author=. Journal of Classification , volume=. 1985 , publisher=

    1985

  16. [24]

    Statistics and Computing , volume=

    Model-based clustering of multiple networks with a hierarchical algorithm , author=. Statistics and Computing , volume=. 2024 , publisher=

    2024

  17. [25]

    Biometrika , volume=

    Estimating network edge probabilities by neighbourhood smoothing , author=. Biometrika , volume=. 2017 , publisher=

    2017

  18. [26]

    Statistics and Computing , volume=

    A tutorial on spectral clustering , author=. Statistics and Computing , volume=. 2007 , publisher=

    2007

  19. [27]

    2013 , publisher=

    Matrix Analysis , author=. 2013 , publisher=

    2013

  20. [28]

    Lecture Notes on Random Matrix Theory , author=

  21. [29]

    1997 , publisher=

    Handbook of Matrices , author=. 1997 , publisher=

    1997

  22. [30]

    169 of Graduate Texts in Mathematics , author=

    Matrix Analysis, Vol. 169 of Graduate Texts in Mathematics , author=. 1997 , publisher=

    1997

  23. [31]

    Tao, Terence , year=

  24. [32]

    Advances in Neural Information Processing Systems , volume=

    Noise thresholds for spectral clustering , author=. Advances in Neural Information Processing Systems , volume=

  25. [33]

    arXiv preprint arXiv:1908.01823 , year=

    Change-point detection in dynamic networks via graphon estimation , author=. arXiv preprint arXiv:1908.01823 , year=

    1908 arXiv

  26. [34]

    IEEE Access , volume=

    A Hypothesis Testing for Large Weighted Networks With Applications to Functional Neuroimaging Data , author=. IEEE Access , volume=. 2020 , publisher=

    2020

  27. [35]

    Statistica Sinica , volume=

    A spectral-based framework for hypothesis testing in populations of networks , author=. Statistica Sinica , volume=. 2024 , publisher=

    2024

  28. [36]

    A useful variant of the

    YU, Y and WANG, T and Samworth, RJ , journal=. A useful variant of the

  29. [37]

    The Annals of Applied Statistics , number =

    Subhadeep Paul and Yuguo Chen , title =. The Annals of Applied Statistics , number =. 2020 , doi =

    2020

  30. [38]

    The Annals of Statistics , volume=

    Global and individualized community detection in inhomogeneous multilayer networks , author=. The Annals of Statistics , volume=. 2022 , publisher=

    2022

  31. [39]

    Electronic Journal of Statistics , volume=

    Estimating a network from multiple noisy realizations , author=. Electronic Journal of Statistics , volume=

  32. [40]

    arXiv preprint arXiv:2107.03431 , year=

    Bayesian model-based clustering for multiple network data , author=. arXiv preprint arXiv:2107.03431 , year=

    arXiv

  33. [41]

    Physical Review E , volume=

    Clustering of heterogeneous populations of networks , author=. Physical Review E , volume=. 2022 , publisher=

    2022

  34. [42]

    2021 55th Asilomar Conference on Signals, Systems, and Computers , pages=

    Network recovery from unlabeled noisy samples , author=. 2021 55th Asilomar Conference on Signals, Systems, and Computers , pages=

    2021

  35. [43]

    IEEE transactions on network science and engineering , volume=

    Clustering network layers with the strata multilayer stochastic block model , author=. IEEE transactions on network science and engineering , volume=. 2016 , publisher=

    2016

  36. [44]

    The Annals of Statistics , volume=

    Community detection on mixture multilayer networks via regularized tensor decomposition , author=. The Annals of Statistics , volume=. 2021 , publisher=

    2021

  37. [45]

    Statistical Modelling , volume=

    Model-based clustering for populations of networks , author=. Statistical Modelling , volume=. 2020 , publisher=

    2020

  38. [46]

    arXiv preprint arXiv:1606.05277 , year=

    Stochastic blockmodels for exchangeable collections of networks , author=. arXiv preprint arXiv:1606.05277 , year=

    arXiv

  39. [47]

    Statistical modelling , volume=

    Analysis of association football playing styles: An innovative method to cluster networks , author=. Statistical modelling , volume=. 2019 , publisher=

    2019

  40. [48]

    The Annals of Applied Statistics , volume=

    Communication network dynamics in a large organizational hierarchy , author=. The Annals of Applied Statistics , volume=. 2024 , publisher=

    2024

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.