Pith. sign in

REVIEW 4 major objections 7 minor 38 references

Bayesian Gaussian Mixture Modeling for Symmetric Matrix Variate Data

T0 review · 4 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Mixture model reveals crime and youth jobs drive mobility overlap

desk verdict A genuinely useful reduced-order parameterization for symmetric matrix-variate precision matrices, but the m-selection evidence is thinner than the application's universal m=1 claim suggests. read the letter →

arxiv 2608.06658 v1 pith:RLCIFFCF submitted 2026-08-07 stat.ME

Malcolm Wolff , Adrian Dobra , Anton H. Westveld , Grace S. Chiu This is my paper · ORCID
classification stat.ME MSC 62H1262F1562H30
keywords symmetricmatrix-variatenormalGaussianmixturemodelreversible-jumpMCMCG-Wishartpriorprecisionmatrixpolynomialhumanmobilitysociomatrixcommutativityconstraint
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 sets out to show that the commutativity constraint inherent in the symmetric matrix-variate normal distribution can be converted into a large reduction in parameters, making it possible to infer relationships between networks from very few observed matrices. The practical target is individual-level mobility data, which arrive as a short series of large, sparse, symmetric sociomatrices. The model called STRUCTURED writes each component's column precision matrix as a polynomial of its row precision matrix, and its reversible-jump variant, STRUCTURED-RJ, selects a single active term of the form $\Upsilon = aI + b\Omega^m$, cutting the effective parameter count to three scalars per component. Simulation results indicate that this reduced parameterization beats unconstrained estimation when only a handful of observations per component are available. In the King County application, the model assigns 6 of 11 weeks to youth employment and 5 to crime incidence, with $m=1$ selected for every component, meaning the same demographic graph explains both the 'who' and the 'how' of activity overlap.

What carries the argument

The load-bearing object is the commutativity constraint $\Omega\Upsilon = \Upsilon\Omega$ in the symmetric matrix-variate normal distribution. Commutativity makes the two precision matrices simultaneously diagonalizable, so $\Upsilon$ can be expressed as a polynomial of $\Omega$; the model then restricts attention to a single active term $\Upsilon = aI + b\Omega^m$, with the polynomial order $m$ treated as an unknown parameter and selected by reversible-jump MCMC. The likelihood's normalizing constant is evaluated through the eigendecomposition of $\Omega$, avoiding explicit construction of the $p^2 \times p(p+1)/2$ transformation matrix $B_p$ and making the sampler computationally tractable.

What would settle it

Simulate 11 weekly sociomatrices from a symmetric matrix-variate normal mixture with non-commuting precision matrices and compare STRUCTURED-RJ's joint precision MSE to the unconstrained MVGGM; if the reduced model's MSE does not at least match the unconstrained model, the commuting assumption is doing the work.

Watch

Extended reading notes

Core claim

The central claim is that for symmetric matrix-variate normal data, commutativity of the row and column precision matrices forces the column precision to be a polynomial of the row precision, and that the reduced form $\Upsilon = aI + b\Omega^m$ with order $m$ chosen by reversible-jump MCMC captures that polynomial structure with three effective parameters. The paper argues that this reduction is what makes estimation feasible in sparse-data regimes, where an unconstrained estimator must fit $p(p+1)$ free elements per precision matrix. Simulation comparisons show STRUCTURED-RJ achieving roughly ten times lower joint-precision MSE than the unconstrained MVGGM when $\Upsilon = \Omega$, and outperforming all alternatives with as few as five observations per component. The application to GPS-derived sociomatrices of 293 King County residents identifies crime incidence and youth employment as the demographic similarity graphs that best explain weekly variation in activity overlap, with every component selecting $m=1$ and thus a shared conditional-independence graph for row and column precision.

Load-bearing premise

The model's conclusions stand only if the true row and column precision matrices commute and the column precision is exactly one active polynomial term; if either fails, the reduced parameterization is misspecified.

Editorial extensions

If this is right

  • With as few as five observations per mixture component, the reduced-order parameterization delivers lower MSE than the unconstrained MVGGM, so short network time series become analyzable.
  • Automatic selection of $m$ by reversible-jump MCMC costs almost nothing: STRUCTURED-RJ stays within about 5% of the oracle fixed-order model for $m=1$, so practitioners do not need to know the polynomial order in advance.
  • For larger samples (about ten or more observations per component), the full-polynomial variant STRUCTURED-FP is the better choice, especially when the true relation has degree $m=2$.
  • In the King County data, the same demographic similarity graph governs both row and column precision ($m=1$ for all components), which simplifies interpretation of activity-overlap networks.
  • The low-earnings employment graph is effectively inactive ($\hat{\pi}=0.07$), suggesting that local crime and youth employment density, not low-wage job density, are the demographic dimensions that differentiate weekly mobility overlap.

Reading between the lines

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

  • If the commuting-precision assumption is plausible, the same parameter reduction transfers to other symmetric relational datasets with few temporal replicates, such as financial correlation networks or protein interaction snapshots.
  • The reported crime/youth split depends on threshold-dichotomized demographic similarity graphs; replacing those thresholds with a fully Bayesian graph-selection prior would test whether the split is an artifact of the discretization.
  • Adding temporal persistence across weeks, for instance a hidden Markov model on component labels, could reveal mobility regimes that last longer than a single week and sharpen the demographic interpretation.
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

4 major / 7 minor

Summary. The paper introduces STRUCTURED, a Bayesian mixture model for symmetric matrix-variate normal data. The key idea is to exploit the commutativity of the row and column precision matrices, Ω and Υ, to express Υ as a polynomial of Ω, and then to restrict the model to a single active polynomial term Υ = aI + bΩ^m, with the order m selected by reversible-jump MCMC (STRUCTURED-RJ) or a full polynomial (STRUCTURED-FP). Simulation studies claim that STRUCTURED-RJ outperforms the unconstrained MVGGM in sparse-data regimes, while STRUCTURED-FP is preferred for larger samples. The method is applied to weekly GPS-derived sociomatrices of individuals in King County, WA, where the mixture components are tied to demographic similarity graphs built from crime incidence and employment data, and the model assigns 5 weeks to crime and 6 weeks to youth employment, with m = 1 selected for all components.

Significance. If the theoretical and empirical claims were fully supported, the reduced parameterization for symmetric matrix-variate normal mixtures would be a useful contribution for high-dimensional sparse relational data. The paper is original in combining the symmetric matrix-variate normal distribution with a mixture over demographic graphs and in using reversible-jump MCMC to select the polynomial order. The application to GPS-derived activity overlap is timely and the comparison with MVGGM is a reasonable external baseline. However, several load-bearing points are not adequately established: the polynomial representation theorem is overstated, the reversible-jump procedure's ability to recover m = 2 is not demonstrated and the reported simulations suggest it may fail, and the application's m = 1 finding is likely dominated by the prior rather than the data. These gaps currently limit confidence in the central claim of accurate reduced-order estimation.

major comments (4)
  1. [Section 3, Theorem 3.1 and Appendix A.2] Theorem 3.1 claims that for any commuting Υ and Ω, Υ is a limit of polynomials in Ω + ε, even when Ω has repeated eigenvalues. This statement is not correct in general. For example, if Ω = I, then every polynomial in Ω + ε is scalar, so no limit of such polynomials can equal a non-scalar commuting Υ. The proof in Appendix A.2 approximates Ω by a simple-spectrum matrix but does not control the coefficients α_k(ε), and the limit need not be a polynomial in Ω. Since the polynomial parameterization is the methodological foundation, the authors should either restrict the theorem to the almost-sure case of simple eigenvalues and explicitly state the reduced form as a modeling assumption for the general case, or provide a correct proof.
  2. [Section 4, Tables 1 and 3] The paper never reports the posterior distribution of m for the simulation with true Υ = Ω^2. Table 1 reports P(m = 1) ≈ 0.84 for true m = 1, but for true m = 2 only MSEs are given. Table 3 shows that STRUCTURED-RJ's MSE for m = 2 is essentially flat (0.044, 0.045, 0.047, 0.052) as N increases from 5 to 50, whereas MVGGM improves from 0.145 to 0.029. If the sampler were correctly selecting m = 2, increasing N should reduce the RJ estimator's error; the flat profile instead suggests that the sampler frequently selects the prior-favored m = 1, imposing a misspecified linear structure. This is a load-bearing gap because the central claim that STRUCTURED-RJ accurately recovers the reduced-order parameterization depends on the sampler honestly identifying m.
  3. [Section 5, Table 4 and Figure 6] The application's universal posterior probability of 1.0 for m = 1 is not informative about the true polynomial order. With only N/L ≈ 4 effective observations per component and the geometric prior p(m) ∝ ρ^m, the prior will dominate the posterior, as the authors themselves note that the simulation advantage of STRUCTURED-RJ is confined to N ≤ 5 per component. The active components have 5 and 6 weeks, which is near the crossover where MVGGM and STRUCTURED-FP become competitive (Table 3), so the regime claim is fragile. The interpretation that 'Corr(Ω̂, Υ̂) = 1.00 confirms that the linear relationship Υ ≈ bΩ is empirically valid' is therefore not supported; it may merely reflect the prior.
  4. [Section 3, Eq. (13)] The normalizing constant formula in Eq. (13) is asserted with a proportionality sign and no derivation. The determinant of B_p^T(Ω^{-1} ⊗ Υ^{-1})B_p does factor into a product of (σ_i ψ_j + σ_j ψ_i) terms for commuting Ω and Υ, but with a factor of 2^{-p} that is omitted by the i ≤ j product as written. While this factor is constant and thus cancels in Metropolis-Hastings ratios, the exact relationship to Eq. (4) should be stated and proved in Appendix A.2. If any parameter-dependent factor had been omitted, the MCMC would be invalid, so this needs to be made precise.
minor comments (7)
  1. [Section 5, first paragraph vs. '40 actors'] The application text first states that 293 individuals are used, but later says 'among 40 actors'. This discrepancy should be clarified, and the reason for the reduction to 40 actors should be explained.
  2. [Section 4, Table 2] In the mixed-polynomial case (0.5Ω + 0.5Ω^2), STRUCTURED-RJ achieves a lower MSE than the oracle FIXED model (0.011 vs. 0.015), even though the single-term restriction is misspecified. A brief explanation of why the restricted model outperforms the oracle in this setting would be helpful.
  3. [Section 3, Eq. (12)] The reversible-jump acceptance ratio does not explicitly state the proposal densities for the reverse move. Since the continuous parameter space has the same dimension for all m, no Jacobian is needed, but this should be stated for clarity.
  4. [Section 3.1] The parallel tempering scheme is described only briefly; the number of chains, temperature schedule, and swap frequency are not given. These details are needed for reproducibility.
  5. [Section 5, Table 4] The low-earnings component is reported as 'effectively inactive' with '—' for sparsity. It would be clearer to report the posterior mean of the precision matrices for each component, including those with small π, or to state why they are omitted.
  6. [References] The Anderson (1958) reference is listed as 'Absence of diffusion in certain random lattices', which is a physics paper, not the usual Anderson (1958) reference for Wishart matrices. Please check the citation.
  7. [Section 5, Table 5] The diagnostic comparison between CPT and LFCM is relevant, but the mixture entropy for both is nearly identical (0.89 vs. 0.90); the text's claim that LFCM 'allows the model to differentiate between weeks' would be strengthened by additional diagnostics, such as posterior assignment entropy per week.

Circularity Check

1 steps flagged · score 6.0 of 10

The claim that Corr(Ω̂,Υ̂)=1.00 confirms the linear relationship Υ≈bΩ is a self-definitional artifact of the model; the core simulation comparisons remain externally grounded.

  1. self definitional [Section 5, 'Posterior Estimates' (Table 4 and following paragraph)]
    "The reversible-jump sampler selects m=1 for all three components, and the near-unity correlations Corr(Ω̂_l, Υ̂_l)=1.00 for the active components confirm that the linear relationship Υ_l ≈ b_l Ω_l is empirically valid."

    Under the model in Eq. (10), Υ_l = a_l I + b_l Ω_l^{m_l}. When m_l=1 is selected, the off-diagonal entries of Υ_l are exactly b_l times the off-diagonal entries of Ω_l. Therefore any nonzero b_l makes the correlation between the off-diagonal elements of Ω̂_l and Υ̂_l identically 1, regardless of the data. The paper presents this correlation as empirical confirmation that the linear relationship holds, but it is an algebraic consequence of the parameterization, not an independent finding. The 'empirical validity' claim is thus equivalent to the model's own definition.

full rationale

The paper's central methodological derivation is not circular: Theorem 3.1 follows from simultaneous diagonalizability (Lemma A.3) and is standard linear algebra; Eq. (9) is explicitly introduced as a restriction ('We address both issues by restricting to a single active polynomial term'), not as a derived prediction; and the simulation comparisons against MVGGM, FIXED, and SUM-K use independent data-generating processes, so the claimed sparse-data advantage is externally grounded. The application's use of the authors' LFCM (Wolff et al., 2025) is a self-citation, but it is a preprocessing dependency acknowledged in the limitations ('they inherit the assumptions of the LFCM model itself'), not a step where the target result is assumed. The one genuine circular step is in Section 5: the paper states that Corr(Ω̂_l, Υ̂_l)=1.00 'confirm[s] that the linear relationship Υ_l ≈ b_l Ω_l is empirically valid.' Under the fitted model (10) with m_l=1, Υ_l is defined as a_l I + b_l Ω_l, so the off-diagonal correlation is exactly 1 by construction for any fitted values; this 'confirmation' reduces to the model's own definition and carries no evidential weight. The related claim that unanimous m=1 selection 'provides empirical support' is weakened by the geometric prior p(m)∝ρ^m and N/L≈4, but that is a prior-sensitivity limitation rather than a definitional circularity. Because this circular confirmation is used to validate the application's substantive interpretation, while the core estimation methodology has independent simulation support, the paper is partially circular (score 6).

Assumptions & free parameters 5 free parameters · 7 assumptions · 0 invented entities

The paper relies on a chain of modeling assumptions: the symmetric matrix-variate normal with commuting precisions, the single-power restriction, known demographic graphs, and the self-cited LFCM preprocessing. The free parameters include per-component scale and order terms, the threshold that builds the graphs, the fixed number of components, and an unexplained 40-actor subset. No entirely new physical or mathematical entities are introduced.

free parameters (5)
  • a_l and b_l (per-component scales) = estimated by MCMC; not reported numerically
    Each mixture component l has a non-negative intercept a_l and positive slope b_l in Υ=a_l I + b_l Ω^m (Eq 9); these are learned from the data with Exponential priors.
  • m_l (polynomial order) = 1 in the application for all components
    The reversible-jump move proposes m±1 and the posterior concentrates on m=1; the geometric prior p(m)∝ρ^m favors low orders.
  • τ (similarity threshold) = 15th percentile of pairwise exposure differences
    Dichotomizes continuous exposure rates into the demographic graphs G_l; the choice determines graph sparsity (174-234 edges) and is not justified beyond convention.
  • L (number of mixture components) = 3
    The number of components is fixed to the three demographic similarity graphs; the paper acknowledges extension to unknown L as future work.
  • Number of actors used in application = 40 (from 293 devices)
    The analysis uses only 40 actors without explaining the subset rule, which affects all downstream graphs and assignments.
assumptions (7)
  • domain assumption The symmetric matrix-variate normal distribution (Nel 1978) is a valid generative model for the sociomatrices.
    Definition 2.2 is invoked throughout; it requires ΩΥ=ΥΩ and continuous support, which LFCM sociomatrices approximate only with 3% near-zero entries.
  • domain assumption The row and column precision matrices commute: ΩΥ=ΥΩ.
    This is the defining constraint that enables the polynomial parameterization; it is assumed, not tested, and underlies every likelihood evaluation (Eq 4).
  • standard math Any matrix commuting with a diagonalizable Ω with simple spectrum is a polynomial of Ω (Lemma 3.1).
    Standard linear algebra; the paper extends it via an epsilon-limit (Theorem 3.1) that is not fully justified for repeated eigenvalues.
  • ad hoc to paper The single-active-term restriction Υ=a I + b Ω^m is adequate for the data.
    Eq (9) reduces the polynomial to one term; the paper's own simulations include a two-term truth (0.5Ω+0.5Ω^2) where the reduced model still performs reasonably, but no justification is given for why real data should satisfy this form.
  • domain assumption The demographic similarity graphs G_l are known and correctly encode the conditional independence structure.
    Section 3 states the graphs are known; in the application they are constructed from crime, youth employment, and low earnings via thresholding, so the known graph is really a fitted object.
  • ad hoc to paper The LFCM trajectories faithfully represent individual activity overlap.
    The sociomatrices are built from the authors' own Lévy flight cluster model (Wolff et al. 2025); the diagnostics in Table 5 are computed on the same data and are not an independent validation.
  • domain assumption The G-Wishart prior with the demographic graph is an appropriate prior for Ω_l.
    Definition 2.3; the posterior sparsity (80-87%) is a direct consequence of the graph structure imposed by the prior.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bayesian Gaussian Mixture Modeling for Symmetric Matrix Variate Data." pith.science (2026). https://pith.science/paper/RLCIFFCF

@misc{pith2026260806658,
  author       = {Pith},
  title        = {Pith review of: Bayesian Gaussian Mixture Modeling for Symmetric Matrix Variate Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RLCIFFCF}},
  note         = {Machine review of arXiv:2608.06658}
}
read the original abstract

Statistical inference on individual activity networks has been a historically difficult task due to the lack of available data at the appropriate granularity and the complexity of modeling individual mobility patterns. The recent availability of GPS data from individual devices, combined with highly detailed demographic information, suggests that one of these challenges can now be addressed. We introduce a new model which we call the Symmetric Matrix-Variate Normal Mixture Model (STRUCTURED) to estimate how demographic traits influence changes in human activity networks, using sociomatrices that capture the probabilistic spatial overlap between individuals over time. We exploit the commutativity constraint inherent in the symmetric matrix-variate normal distribution to parameterize the column precision matrix as a polynomial of the row precision matrix, reducing the effective parameter space by an order of magnitude. We develop two variants of STRUCTURED: STRUCTURED-FP, which estimates the full polynomial, and STRUCTURED-RJ, which uses reversible-jump MCMC to select a reduced-order parameterization. Simulation studies demonstrate that STRUCTURED-RJ outperforms existing methods in sparse-data regimes, whereas STRUCTURED-FP is preferred when sample sizes are large. We apply the model to GPS-derived sociomatrices of 293 individuals in King County, WA, finding that local crime environments and youth employment density are the dominant demographic factors explaining variation in weekly activity overlap patterns.

Figures

Figures reproduced from arXiv: 2608.06658 by the authors.

Figure 1
Figure 1. Cumulative average sum of matrix elements for Ω (left panel) and Υ (right panel) by iteration across 10 independent chains. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. displays the estimated activity distribution for an example individual in a given week; latitude and longitude are normalized to preserve privacy. Although a fair amount of spread is observed across the space, the individual spends the majority of time in the week around (-1,1) [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. displays the distributions of pairwise log distances between individuals by week, where the distances are computed using Eq. (15). Colors indicate the number of missing devices in that particular week of the 293 total devices. Fifty devices were not observed in the first week, a substantially higher number than in the remaining weeks, ranging from 0– 25 missing devices. Moreover, the observed pairwise distance distr… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: displays the resulting sociomatrix for a representative week, where entry pi, jq rep￾resents the average inverse L2 distance between the generated trajectories of actors i and j [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: displays the posterior mean estimates of Ωl and Υl for each component. The reversible-jump sampler selects m “ 1 for all three components with posterior probability 1.0, indicating that the linear relationship Υl « blΩl provides the best fit in this sparse-data setting…
Figure 6
Figure 6. Figure 6: Posterior mean probability of assignment to each graph across 11 weeks. that the linear relationship Υl « blΩl is empirically valid. The sparsity of both precision ma￾trices (80–87% of off-diagonal elements below 10´3 ) reflects the sparse nature of the underlying demo…
Figure 7
Figure 7. Figure 7: Convergence diagnostics for the application. Left panel: cumulative mean of mixing weights πl . Right panel: cumulative fraction of weeks assigned to each component. The LFCM-based sociomatrices are preferable for the symmetric matrix-variate normal model on two ground…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 35 canonical work pages

  1. [1]

    Absence of diffusion in certain random lattices

    Philip W Anderson. Absence of diffusion in certain random lattices. Physical Review, 109 0 (5): 0 1492, 1958

    work page 1958

  2. [2]

    Anderson

    Theodore W. Anderson. An Introduction to Multivariate Statistical Analysis. John Wiley & Sons, Hoboken, NJ, 3rd edition, 2003. ISBN 978-0-471-36091-9

    work page 2003

  3. [3]

    A variable order M arkov model approach for mobility prediction

    Halgurt Bapierre, Georg Groh, and Stefan Theiner. A variable order M arkov model approach for mobility prediction. Pervasive Computing, pages 8--16, 2011

    work page 2011

  4. [4]

    Uncovering the socioeconomic facets of human mobility

    Hugo Barbosa, Surendra Hazarie, Brian Dickinson, Aleix Bassolas, Adam Frank, Henry Kautz, Adam Sadilek, Jos \'e J Ramasco, and Gourab Ghoshal. Uncovering the socioeconomic facets of human mobility. Scientific reports, 11 0 (1): 0 1--13, 2021

    work page 2021

  5. [5]

    Modeling covariance matrices in terms of standard deviations and correlations, with application to shrinkage

    John Barnard, Robert McCulloch, and Xiao-Li Meng. Modeling covariance matrices in terms of standard deviations and correlations, with application to shrinkage. Statistica Sinica, pages 1281--1311, 2000

    work page 2000

  6. [6]

    Exploring sociocentric and egocentric approaches for social network analysis

    Kenneth KS Chung, Liaquat Hossain, and Joseph Davis. Exploring sociocentric and egocentric approaches for social network analysis. In Proceedings of the 2nd international conference on knowledge management in Asia Pacific, pages 1--8, 2005

    work page 2005

  7. [7]

    Crime rate and social integration the impact of metropolitan mobility

    Robert D Crutchfield, Michael R Geerken, and Walter R Gove. Crime rate and social integration the impact of metropolitan mobility. Criminology, 20 0 (3-4): 0 467--478, 1982

    work page 1982

  8. [8]

    Covariance selection

    Arthur P Dempster. Covariance selection. Biometrics, pages 157--175, 1972

    work page 1972

Show all 38 references
  1. [9]

    Eigenvectors from eigenvalues: a survey of a basic identity in linear algebra

    Peter Denton, Stephen Parke, Terence Tao, and Xining Zhang. Eigenvectors from eigenvalues: a survey of a basic identity in linear algebra. Bulletin of the American Mathematical Society, 59 0 (1): 0 31--58, 2022

    2022

  2. [10]

    Bayesian inference for general gaussian graphical models with application to multivariate lattice data

    Adrian Dobra, Alex Lenkoski, and Abel Rodriguez. Bayesian inference for general gaussian graphical models with application to multivariate lattice data. Journal of the American Statistical Association, 106 0 (496): 0 1418--1433, 2011

    2011

  3. [11]

    A statistical framework for measuring the temporal stability of human mobility patterns

    Zhihang Dong, Yen-Chi Chen, and Adrian Dobra. A statistical framework for measuring the temporal stability of human mobility patterns. Journal of Applied Statistics, pages 1--19, 2020

    2020

  4. [12]

    Gender gaps in urban mobility

    Laetitia Gauvin, Michele Tizzoni, Simone Piaggesi, Andrew Young, Natalia Adler, Stefaan Verhulst, Leo Ferres, and Ciro Cattuto. Gender gaps in urban mobility. Humanities and Social Sciences Communications, 7 0 (1): 0 1--13, 2020

    2020

  5. [13]

    Matrix variate distributions, volume 104

    Arjun K Gupta and Daya K Nagar. Matrix variate distributions, volume 104. CRC Press, 2018

    2018

  6. [14]

    Discrete temporal models of social networks

    Steve Hanneke, Wenjie Fu, and Eric Xing. Discrete temporal models of social networks. arXiv preprint arXiv:0908.1258, 2009

    2009 arXiv

  7. [15]

    Modeling dependencies in international relations networks

    Peter D Hoff and Michael D Ward. Modeling dependencies in international relations networks. Political Analysis, pages 160--175, 2004

    2004

  8. [16]

    Latent space approaches to social network analysis

    Peter D Hoff, Adrian E Raftery, and Mark S Handcock. Latent space approaches to social network analysis. Journal of the American Statistical Association, 97 0 (460): 0 1090--1098, 2002

    2002

  9. [17]

    Bayes factors

    Robert E Kass and Adrian E Raftery. Bayes factors. Journal of the American Statistical Association, 90 0 (430): 0 773--795, 1995

    1995

  10. [18]

    The reversal small-world experiment

    Peter D Killworth and H Russell Bernard. The reversal small-world experiment. Social networks, 1 0 (2): 0 159--192, 1978

    1978

  11. [19]

    Acquaintance networks between racial groups: Application of the small world method

    Charles Korte and Stanley Milgram. Acquaintance networks between racial groups: Application of the small world method. Journal of Personality and Social Psychology, 15 0 (2): 0 101, 1970

    1970

  12. [20]

    Bayesian inference for a covariance matrix

    Tom Leonard and John SJ Hsu. Bayesian inference for a covariance matrix. The Annals of Statistics, 20 0 (4): 0 1669--1696, 1992

    1992

  13. [21]

    Bayesian correlation estimation

    John C Liechty, Merrill W Liechty, and Peter M \"u ller. Bayesian correlation estimation. Biometrika, 91 0 (1): 0 1--14, 2004

    2004

  14. [22]

    How peer influence affects attribute preferences: A bayesian updating mechanism

    Vishal Narayan, Vithala R Rao, and Carolyne Saunders. How peer influence affects attribute preferences: A bayesian updating mechanism. Marketing Science, 30 0 (2): 0 368--384, 2011

    2011

  15. [23]

    D.G. Nel. On the symmetric multivariate normal distribution and the asymptotic expansion of a wishart matrix. South African Statistical Journal, 12 0 (2): 0 145--159, 1978

    1978

  16. [24]

    Geographic constraints on social network groups

    Jukka-Pekka Onnela, Samuel Arbesman, Marta C Gonz \'a lez, Albert-L \'a szl \'o Barab \'a si, and Nicholas A Christakis. Geographic constraints on social network groups. PLoS one, 6 0 (4): 0 e16939, 2011

    2011

  17. [25]

    A symmetric matrix-variate normal local approximation for the wishart distribution and some applications

    Fr \'e d \'e ric Ouimet. A symmetric matrix-variate normal local approximation for the wishart distribution and some applications. Journal of Multivariate Analysis, 189: 0 104923, 2022

    2022

  18. [26]

    A framework for criminal network analysis using big data

    Md Ileas Pramanik, Wenping Zhang, Raymond YK Lau, and Chunping Li. A framework for criminal network analysis using big data. In 2016 IEEE 13th international conference on e-business engineering (ICEBE), pages 17--23. IEEE, 2016

    2016

  19. [27]

    Flexible covariance estimation in graphical G aussian models

    Bala Rajaratnam, H \'e l \`e ne Massam, and Carlos M Carvalho. Flexible covariance estimation in graphical G aussian models. The Annals of Statistics, 36 0 (6): 0 2818--2849, 2008

    2008

  20. [28]

    Analysing long term spatial mobility patterns of individuals and large groups using 3d-gis: A sport geographic approach

    Sebastian Rauch. Analysing long term spatial mobility patterns of individuals and large groups using 3d-gis: A sport geographic approach. Tijdschrift voor economische en sociale geografie, 2022

    2022

  21. [29]

    Hyper inverse W ishart distribution for non-decomposable graphs and its application to B ayesian inference for G aussian graphical models

    Alberto Roverato. Hyper inverse W ishart distribution for non-decomposable graphs and its application to B ayesian inference for G aussian graphical models. Scandinavian Journal of Statistics, 29 0 (3): 0 391--411, 2002

    2002

  22. [30]

    Bayesian model selection for the latent position cluster model for social networks

    Caitriona Ryan, Jason Wyse, and Nial Friel. Bayesian model selection for the latent position cluster model for social networks. Network Science, 5: 0 70--91, 03 2017. doi:10.1017/nws.2017.6

    2017 doi

  23. [31]

    A parallel tempering algorithm for probabilistic sampling and multimodal optimization

    Malcolm Sambridge. A parallel tempering algorithm for probabilistic sampling and multimodal optimization. Geophysical Journal International, 196 0 (1): 0 357--374, 2014

    2014

  24. [32]

    Socio-spatial properties of online location-based social networks

    Salvatore Scellato, Anastasios Noulas, Renaud Lambiotte, and Cecilia Mascolo. Socio-spatial properties of online location-based social networks. In Proceedings of the International AAAI Conference on Web and Social Media, volume 5, pages 329--336, 2011

    2011

  25. [33]

    Parsimonious covariance matrix estimation for longitudinal data

    Michael Smith and Robert Kohn. Parsimonious covariance matrix estimation for longitudinal data. Journal of the American Statistical Association, 97 0 (460): 0 1141--1153, 2002

    2002

  26. [34]

    Social networks and industrial geography

    Olav Sorenson and Pino G Audia. Social networks and industrial geography. Journal of Evolutionary Economics, 13 0 (5): 0 513--527, 2003

    2003

  27. [35]

    An experimental study of the small world problem

    Jeffrey Travers and Stanley Milgram. An experimental study of the small world problem. In Social Networks, pages 179--197. Elsevier, 1977

    1977

  28. [36]

    Bayesian graphical lasso models and efficient posterior computation

    Hao Wang. Bayesian graphical lasso models and efficient posterior computation. Bayesian Analysis, 7 0 (4): 0 867--886, 2012

    2012

  29. [37]

    Westveld, and Grace S

    Malcolm Wolff, Adrian Dobra, Anton H. Westveld, and Grace S. Chiu. Modeling human spatial mobility patterns with the L \'evy flight cluster model, 2025. URL https://arxiv.org/abs/2509.00298

    2025 arXiv

  30. [38]

    Estimation of a covariance matrix using the reference prior

    Ruoyong Yang and James O Berger. Estimation of a covariance matrix using the reference prior. The Annals of Statistics, pages 1195--1211, 1994

    1994

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.