Pith. sign in

REVIEW 2 major objections 6 minor 197 references

Fast Stochastic Nearest Neighbor Pairwise Composite Likelihood for Massive Spatial Datasets

T0 review · 2 major / 6 minor · reviewed 2026-07-08 · glm-5.2

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

Pith's one-line read Two pairs per observation suffice for massive spatial covariance estimation

desk verdict Stochastic thinning of NN pairwise composite likelihoods for massive spatial data — solid practical contribution with honest empirical validation, but the WorldClim prediction comparison is less discriminating than the paper implies. read the letter →

arxiv 2607.06142 v1 pith:ZUJJYURM submitted 2026-07-07 stat.ME stat.CO

Moreno Bevilacqua , Francisco Cuevas-Pacheco , Christian Caamaño-Carrillo This is my paper
classification stat.MEstat.CO
keywords stochasticcompositelikelihoodnearest-neighbormethodsspatialstatisticsMatérncovariancelargedatasetsthinningVecchiaapproximationpairwise
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

This paper proposes a stochastic acceleration of nearest-neighbor pairwise composite likelihood for spatial Gaussian random fields. The core idea is to separate two design choices that are coupled in prior work: the size m of the nearest-neighbor candidate graph (which controls the richness of local spatial information) and a thinning parameter p (which controls how many of those candidate pairs are actually evaluated). The authors introduce two thinning designs: Bernoulli thinning, which retains each candidate pair independently with probability p, and fixed-budget thinning, which enforces an exact retained-pair count by allocating a global budget across target-specific neighbor lists and sampling without replacement within each. The thinned composite likelihood is then maximized as usual. Under standard increasing-domain asymptotics, the resulting estimators are consistent and asymptotically normal with a Godambe sandwich covariance, and standard errors can be estimated via a parametric score bootstrap that reuses the precomputed neighbor graph. The central empirical finding is that retaining approximately two pairs per observation (K_tar/n = 2) provides a stable statistical-computational compromise for Matérn covariance models: the mean, scale, and sill parameters remain essentially as efficient as the full nearest-neighbor estimator, while the main efficiency loss is concentrated on the smoothness parameter. On a 2.5-million-observation WorldClim temperature dataset, the method achieves predictive accuracy indistinguishable from a Vecchia benchmark at 10-25x shorter covariance-fitting time.

What carries the argument

The central mechanism is the two-stage decomposition of the nearest-neighbor pairwise composite likelihood: (1) construct a deterministic directed m-nearest-neighbor candidate graph on the observation locations, and (2) evaluate only a randomized subset of its pairwise likelihood contributions. The thinning parameter p (or equivalently the retained-pair budget K_tar/n) controls the computational cost independently of m. This separation allows using a rich candidate graph (large m) while keeping the number of evaluated bivariate likelihood terms at O(n), specifically approximately 2n terms at the calibrated budget.

What would settle it

A spatial dataset with strongly non-uniform sampling (e.g., clustered monitoring stations) and an anisotropic or non-Matérn covariance where the K_tar/n = 2 budget produces substantially degraded estimates of the scale or sill parameters — not just smoothness — would undermine the claim that two pairs per observation is a broadly stable compromise.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the deterministic nearest-neighbor pairwise composite likelihood can be aggressively thinned — retaining only about two pairwise likelihood contributions per observation — while preserving nearly all of the statistical efficiency for mean, variance, and range estimation in Matérn models. The efficiency loss is not uniform across parameters: it is negligible for the mean and sill, moderate for the scale, and largest for the smoothness parameter, where the internal Monte Carlo variability from thinning can account for up to half of the total estimator variance. The two thinning designs (Bernoulli and fixed-budget) behave similarly at the same budget; the固定

Load-bearing premise

The empirical budget of two retained pairs per observation is calibrated exclusively on uniformly distributed sampling locations in two dimensions with stationary isotropic Matérn covariances. The paper itself acknowledges this is not a universal recommendation, but the WorldClim application and all comparative claims against Vecchia rely on this single calibration. Strongly non-uniform sampling, anisotropic covariances, or different covariance families may shift the tradeoff

Editorial extensions

If this is right

  • For practitioners fitting spatial covariance models to millions of observations, the method offers a practical knob: accept a modest loss of smoothness-estimation precision in exchange for order-of-magnitude reductions in covariance-fitting time, while retaining essentially full accuracy for spatial prediction.
  • The thinning framework is defined at the level of the candidate pair graph and is agnostic to the distributional family, so it extends naturally to non-Gaussian, space-time, and multivariate random fields where bivariate likelihoods are available but full likelihoods are intractable.
  • The finding that smoothness is the most sensitive parameter suggests that adaptive thinning — retaining more pairs at spatial lags most informative about smoothness — could recover efficiency for that parameter without proportionally increasing cost.
  • The parametric score bootstrap for Godambe variance estimation, made feasible by the thinning-induced computational savings, provides a route to uncertainty quantification for composite-likelihood spatial fits at scales where it was previously impractical.

Reading between the lines

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

  • If the K_tar/n = 2 budget is specific to uniform sampling and isotropic Matérn covariances, then clustered or preferential sampling designs — common in environmental monitoring networks — may require locally adaptive thinning probabilities that allocate more pairs to under-sampled regions, potentially changing the efficiency-computation trade-off in ways the paper's calibration does not capture.
  • The observation that smoothness estimation is most degraded by thinning, while mean and sill are nearly unaffected, may reflect a fundamental information-geometric asymmetry: smoothness is identified from the curvature of the covariance at short lags, which requires a denser and more geometrically diverse set of local pairs than the other parameters. This suggests a theoretical lower bound on the
  • The fixed-budget design's local negative dependence (negative covariance between indicators sharing the same target) is reminiscent of stratified sampling theory, and one could ask whether optimal allocation of the budget across targets — rather than proportional allocation — could minimize the Godambe variance for a specific parameter of interest.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The paper proposes a stochastic acceleration of nearest-neighbor (NN) weighted pairwise composite likelihoods for massive spatial datasets. The method separates the construction of a deterministic directed NN candidate graph (controlled by m) from the number of pairwise likelihood terms actually evaluated (controlled by a thinning parameter p). Two thinning designs are introduced: Bernoulli thinning, which controls the retained-pair budget in expectation, and fixed-budget thinning, which enforces an exact budget via target-wise sampling without replacement. The paper includes three simulation studies (budget calibration, comparison with a Vecchia approximation, and internal variability) for Matérn covariance models with sample sizes up to 500,000, and a large-scale application to 2.5 million WorldClim temperature observations. An asymptotic sketch for the estimators under increasing-domain conditions is provided in Appendix A.

Significance. The paper addresses a well-defined computational bottleneck in spatial composite likelihood estimation. The separation of the candidate graph richness (m) from the evaluation budget (p) is a natural and useful idea. The simulation studies are thorough, covering up to 500,000 observations with 250 replicates across multiple scenarios. The application to 2.5 million observations demonstrates practical scalability. The implementation in the GeoModels R package and the use of a parametric score bootstrap for Godambe standard errors are practical strengths. The fixed-budget thinning design is a well-motivated contribution that connects to classical finite-population sampling ideas.

major comments (2)
  1. §6, Tables 7–8: The headline claim that the proposed method achieves 'predictive accuracy essentially indistinguishable from the Vecchia benchmark' on the WorldClim data is supported by Table 7 (MAE/RMSPE differences on the order of 10^-4). However, Table 8 shows substantial discrepancies in fitted covariance parameters: Vecchia gives alpha_hat=9.091, sigma^2_hat=2.213, nu_hat=0.831, while the stochastic NN pairwise methods give alpha_hat~11.2–12.1, sigma^2_hat~2.57, nu_hat~0.76–0.79. The paper acknowledges that the GAM mean surface explains ~89% of marginal variation (§6), which likely masks these parameter discrepancies in the prediction metrics. The abstract and conclusion conflate 'predictive accuracy' with overall method adequacy by citing both the simulation results (Table 4, which shows large parameter estimation gaps, especially for smoothness) and the WorldClim prediction metric
  2. §5.1: The empirical budget recommendation K_tar/n = 2 is calibrated exclusively on uniformly distributed sampling locations in 2D with stationary isotropic Matérn covariances. The paper itself acknowledges this is not a universal recommendation (end of §5.1). However, the WorldClim application and all comparative claims against Vecchia rely on this single calibration. The paper would benefit from a brief discussion of how sensitive the K_tar/n = 2 recommendation is to deviations from these assumptions, or at minimum a more prominent caveat in the abstract and conclusion that the budget recommendation is specific to the settings studied.
minor comments (6)
  1. §3.2, Eq. (8): The covariance formula for the fixed-budget design is given conditional on k_j. It would help to clarify whether the marginal covariance (integrating over the randomized rounding of k_j) has a similarly clean form, or whether the conditional expression is the one used in practice.
  2. Table 3: The speedup column reports values like 26.71 and 31.55, but it is not immediately clear from the caption whether these are relative to FullNN in the same scenario. Adding 'Speedup relative to FullNN' to the caption would improve clarity.
  3. §5.2: The comparison between GeoFit and GpGp involves different optimization strategies (direct optimization vs. Fisher scoring, no profiling vs. profiling of mean parameters). The paper acknowledges this, but a brief note in the table caption of Table 4 reminding the reader that these are implementation-level timings would be helpful.
  4. Appendix A, Proposition 1: The asymptotic normality result is stated with K_n^{1/2} normalization. For Bernoulli thinning, K_n is random. The sketch argues K_n/(p d_n) -> 1 in probability, but it would be useful to state explicitly whether the CLT for the score is applied conditionally on the thinning indicators or unconditionally.
  5. Figure 4: The empirical semivariogram is shown only for one configuration (Bernoulli, K_tar/n=1, m=10). Showing the Vecchia fit on the same plot would provide a visual comparison of the covariance parameter discrepancies noted in Table 8.
  6. §4: The parametric score bootstrap avoids refitting the model for each bootstrap dataset. It would be useful to briefly state how many bootstrap replicates were used in the WorldClim application (100 is mentioned in Table 8 caption, but stating it in the main text would help).

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for a careful and constructive report. Both major comments are well-taken and will be addressed in the revision.

read point-by-point responses

  1. Referee: §6, Tables 7–8: The headline claim that the proposed method achieves 'predictive accuracy essentially indistinguishable from the Vecchia benchmark' on the WorldClim data is supported by Table 7 (MAE/RMSPE differences on the order of 10^-4). However, Table 8 shows substantial discrepancies in fitted covariance parameters: Vecchia gives alpha_hat=9.091, sigma^2_hat=2.213, nu_hat=0.831, while the stochastic NN pairwise methods give alpha_hat~11.2–12.1, sigma^2_hat~2.57, nu_hat~0.76–0.79. The paper acknowledges that the GAM mean surface explains ~89% of marginal variation (§6), which likely masks these parameter discrepancies in the prediction metrics. The abstract and conclusion conflate 'predictive accuracy' with overall method adequacy by citing both the simulation results (Table 4, which shows large parameter estimation gaps, especially for smoothness) and the WorldClim prediction metric

    Authors: The referee is correct that the prediction metrics in Table 7 and the covariance parameter estimates in Table 8 tell different stories, and that the GAM mean surface—explaining approximately 89% of marginal variation—substantially masks covariance-parameter discrepancies in the prediction scores. We agree that the current wording in the abstract and conclusion conflates predictive accuracy with broader method adequacy, and we will revise both to make the distinction explicit. Specifically, we will: (1) qualify the WorldClim claim in the abstract to state that *predictive accuracy* is essentially indistinguishable, while noting that fitted covariance parameters differ from the Vecchia benchmark; (2) add a sentence in §6 explicitly acknowledging that the dominant GAM mean surface likely accounts for the near-identical prediction scores despite the parameter differences in Table 8; and (3) ensure the conclusion does not present the simulation parameter-estimation results (Table 4) and the WorldClim prediction results as if they were the same type of evidence. We emphasize that the paper already reports the parameter discrepancies transparently in Table 8 and notes that Vecchia 'optimizes a different approximate likelihood criterion and uses a different fitting workflow,' but we agree the framing in the abstract and conclusion should be tightened. revision: yes

  2. Referee: §5.1: The empirical budget recommendation K_tar/n = 2 is calibrated exclusively on uniformly distributed sampling locations in 2D with stationary isotropic Matérn covariances. The paper itself acknowledges this is not a universal recommendation (end of §5.1). However, the WorldClim application and all comparative claims against Vecchia rely on this single calibration. The paper would benefit from a brief discussion of how sensitive the K_tar/n = 2 recommendation is to deviations from these assumptions, or at minimum a more prominent caveat in the abstract and conclusion that the budget recommendation is specific to the settings studied.

    Authors: The referee is right that the budget recommendation K_tar/n = 2 is calibrated on a specific class of designs (uniform locations in 2D, stationary isotropic Matérn), and that this caveat should be more prominent. The paper already states at the end of §5.1 that the value 'should be interpreted as an empirical calibration for the present class of designs, rather than as a universal recommendation,' and the conclusion discusses how non-uniform sampling, nonstationarity, or different covariance families may require different budgets. However, we agree this qualification is not sufficiently visible in the abstract. We will: (1) add a brief qualifier in the abstract noting that the budget recommendation is specific to the Matérn simulation settings studied; (2) add a short paragraph in §5.1 discussing qualitatively why the recommendation might change under non-uniform sampling (where locally adaptive thinning probabilities could be beneficial) or under covariance models with different short-range behavior; and (3) strengthen the caveat in the conclusion. We note that a full sensitivity analysis across sampling designs and covariance families is beyond the scope of the current paper, but we will make the limitations of the calibration clearer to the reader. revision: yes

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the method is a randomized subsampling of a known objective, and the empirical budget is presented as calibration, not as a derived prediction.

full rationale

The paper proposes stochastic thinning of a deterministic nearest-neighbor pairwise composite likelihood. The core estimator (Eq. 4) is a randomized subsampling of the deterministic NN pairwise objective (Eq. 1-2), targeting the same parameter theta by construction; this is a standard subsampling estimator, not a circular definition. The empirical budget K_tar/n = 2 is calibrated on simulation Study 1 (Section 5.1) and then applied to the WorldClim data (Section 6), but the paper explicitly states this is 'an empirical calibration for the present simulation designs, rather than as a universal recommendation.' No fitted constant is rebranded as a theoretical prediction. The asymptotic theory (Appendix A, Proposition 1) extends standard increasing-domain results for weighted pairwise likelihoods by conditioning on the thinning indicators, which are generated independently of the data; the proof sketch invokes Bevilacqua and Gaetan (2015) for the base case but the extension to thinning is argued directly. Self-citations (Caamaño-Carrillo et al. 2024 for the deterministic NN weights; Bevilacqua et al. 2026 for the GeoModels package and turning-bands simulator) are used as building blocks, not as load-bearing uniqueness theorems that force the paper's conclusion. The Vecchia benchmark is an external method (GpGp package, Guinness 2024). The prediction comparison on WorldClim data may be statistically insensitive to covariance parameter differences (as the skeptic notes, the GAM explains ~89% of variation), but that is a correctness/discriminating-power concern, not a circularity in the derivation. No step in the derivation chain reduces to its own inputs by construction.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No new entities are postulated. The method operates entirely within the existing framework of Gaussian random fields with Matérn covariance and nearest-neighbor weighted pairwise composite likelihoods. The thinning indicators are algorithmic constructs, not new scientific entities.

free parameters (4)
  • m (NN candidate graph size) = 10 (nu=0.5), 50 (nu=1.5)
    Chosen based on smoothness; larger m for smoother fields. Not fitted to data but selected by heuristic.
  • p (thinning fraction) / K_tar/n (retained-pair budget) = K_tar/n = 2 (calibrated)
    Empirically calibrated from Simulation Study 1 as a statistical-computational compromise. Explicitly stated as not a universal recommendation.
  • Vecchia conditioning neighbors = 30
    Fixed for the benchmark; not tuned.
  • GAM basis dimension k = 500
    Chosen for the temperature application mean surface.
assumptions (4)
  • domain assumption Increasing-domain weak-dependence conditions for local weighted pairwise composite likelihood estimation hold (Appendix A, Proposition 1).
    Standard in the spatial composite likelihood literature; invoked to establish consistency and asymptotic normality.
  • domain assumption Thinning indicators are generated independently of the observed field values (Section 3, Eq. 3).
    Structural assumption ensuring the stochastic criterion remains a valid local weighted pairwise likelihood conditional on the thinning realization.
  • domain assumption The retained-pair size K_n satisfies K_n -> infinity and K_n = O(n) (Appendix A, Proposition 1).
    Required for the asymptotic normality result; satisfied by both thinning designs when p is fixed.
  • standard math The normalized sensitivity matrix H is nonsingular in a neighborhood of theta_0 (Appendix A, Proposition 1).
    Standard regularity condition for M-estimator asymptotic normality.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Fast Stochastic Nearest Neighbor Pairwise Composite Likelihood for Massive Spatial Datasets." pith.science (2026). https://pith.science/paper/ZUJJYURM

@misc{pith2026260706142,
  author       = {Pith},
  title        = {Pith review of: Fast Stochastic Nearest Neighbor Pairwise Composite Likelihood for Massive Spatial Datasets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZUJJYURM}},
  note         = {Machine review of arXiv:2607.06142}
}
read the original abstract

Weighted pairwise composite likelihoods based on nearest-neighbor (NN) pairs provide a scalable alternative to full likelihood inference for spatial random fields, but can remain expensive when moderately large NN neighborhoods are needed. We propose a stochastic acceleration that constructs the deterministic NN candidate graph and evaluates only a randomized subset of its pairwise contributions. We consider two thinning designs: Bernoulli thinning, which controls the retained-pair budget in expectation, and fixed-budget thinning, which enforces an exact budget through target-wise sampling without replacement. Simulation studies for Mat\'ern covariance models suggest that, in the settings considered, retaining two pairs per observation provides a stable statistical--computational compromise. The stochastic NN pairwise estimators provide a faster alternative to a Vecchia-type Gaussian approximation when substantial reductions in covariance-fitting time are desired and a modest loss of efficiency is acceptable. This trade-off is especially favorable for the mean, scale, and sill parameters, while the main efficiency loss is concentrated on smoothness estimation. In an application to July average temperature over the western--central United States, based on 2.5 million WorldClim observations, the proposed estimators achieve predictive accuracy essentially indistinguishable from the Vecchia benchmark, with substantially shorter covariance-fitting time.

Figures

Figures reproduced from arXiv: 2607.06142 by the authors.

Figure 1
Figure 1. Toy illustrations of the two stochastic thinning designs on directed nearest-neighbor [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Simulation Study 2. Boxplots of parameter estimates for the largest sample size in the [PITH_FULL_IMAGE:figures/full_fig_p017_2.png] view at source ↗
Figure 3
Figure 3. Simulation Study 2. Boxplots of parameter estimates for the largest sample size in the [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Temperature application. Empirical semivariogram of the GAM residuals and fitted [PITH_FULL_IMAGE:figures/full_fig_p023_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

197 extracted references · 197 canonical work pages

  1. [1]

    , title =

    Cochran, William G. , title =

    work page

  2. [2]

    Sampling Algorithms , publisher =

    Till. Sampling Algorithms , publisher =

    work page

  3. [3]

    Genton , journal =

    Emilio Porcu and Moreno Bevilacqua and Marc G. Genton , journal =. NONSEPARABLE, SPACE-TIME COVARIANCE FUNCTIONS WITH DYNAMICAL COMPACT SUPPORTS , urldate =

    work page

  4. [4]

    ACM Transactions on Mathematical Software (TOMS) , volume=

    Algorithm 1013: An R implementation of a continuous spectral algorithm for simulating vector Gaussian random fields in Euclidean spaces , author=. ACM Transactions on Mathematical Software (TOMS) , volume=. 2020 , publisher=

    work page 2020

  5. [5]

    Composite Likelihood Estimation in Multivariate Data Analysis , urldate =

    Yinshan Zhao and Harry Joe , journal =. Composite Likelihood Estimation in Multivariate Data Analysis , urldate =

    work page

  6. [6]

    Statistics in Biosciences , volume=

    Fast Multivariate Probit Estimation via a Two-Stage Composite Likelihood , author=. Statistics in Biosciences , volume=

    work page

  7. [7]

    Journal of Statistical Software , author=

    On Circulant Embedding for Gaussian Random Fields in R , volume=. Journal of Statistical Software , author=. 2013 , pages=

    work page 2013

  8. [8]

    , title =

    Shinozuka, M. , title =. The Journal of the Acoustical Society of America , year =

    work page

Show all 197 references
  1. [9]

    Simulating space-time random fields with nonseparable

    Allard, Denis and Emery, Xavier and Lacaux, C\'eline and Lantu\'ejoul, Christian , journal=. Simulating space-time random fields with nonseparable. 2020 , publisher=

    2020

  2. [10]

    Geostatistical

    Lantu\'ejoul, Christian , year=. Geostatistical

  3. [11]

    Fast and exact simulation of stationary

    Dietrich, Claude R and Newsam, Garry N , journal=. Fast and exact simulation of stationary. 1997 , publisher=

    1997

  4. [12]

    Andrew T. A. Wood and Grace Chan , title =. Journal of Computational and Graphical Statistics , volume =

  5. [13]

    WIREs Computational Statistics , volume =

    Emilio Porcu and Reinhard Furrer and Douglas Nychka , title =. WIREs Computational Statistics , volume =

  6. [14]

    Porcu and J.M

    E. Porcu and J.M. Montero and M. Schlather and Editors , Publisher =. Advances and Challenges in Space-time Modelling of Natural Events , Year =

  7. [15]

    Gneiting , Journal =

    T. Gneiting , Journal =. Strictly and non-strictly positive definite functions on spheres , Volume =

  8. [16]

    A. O. Finley and A. Datta and B. C. Cook and D. C. Morton and H. E. Andersen and S. Banerjee , Journal =. Efficient Algorithms for Bayesian Nearest Neighbor Gaussian Processes , Volume =

  9. [17]

    Gentz , Journal =

    A. Gentz , Journal =. Numerical computation of multivariate normal probabilities , Volume =

  10. [18]

    Bai and J

    Y. Bai and J. Kang and X. Peter and K. Song , title =. 2014 , volume =

    2014

  11. [19]

    and Hijmans, Robert J

    Fick, Stephen E. and Hijmans, Robert J. , title =. International Journal of Climatology , volume =

  12. [20]

    Hijmans and M

    Robert J. Hijmans and M. 2023 , note =

    2023

  13. [21]

    2013 , Bdsk-Url-1 =

    Douglas Nychka and Reinhard Furrer and Stephan Sain , Date-Added =. 2013 , Bdsk-Url-1 =

    2013

  14. [22]

    Furrer and Marc G

    R. Furrer and Marc G. Genton and D. Nychka , Date-Added =. Covariance tapering for interpolation of large spatial datasets , Volume =. Journal of Computational and Graphical Statistics , Number =

  15. [23]

    Haslett and A

    J. Haslett and A. E. Raftery , Date-Added =. Space-time modelling with long-memory dependence: assessing Ireland's wind power resource , Volume =. Applied Statistics , Number =

  16. [24]

    Gneiting and M

    T. Gneiting and M. Schlather , Date-Added =. Stochastic models that separate fractal dimension and the Hurst effect , Volume =. SIAM Review , Pages =

  17. [25]

    Nonseparable space-time covariance models: some parametric families , Volume =

    Sandra. Nonseparable space-time covariance models: some parametric families , Volume =. Mathematical Geology , Pages =

  18. [26]

    Genton , journal =

    Emilio Porcu and Moreno Bevilacqua and Marc G. Genton , journal =. Nonseparable, space-time covariance functions with dynamical compact supports , volume =

  19. [27]

    Journal of Multivariate Analysis , author =

    Compactly supported correlation functions , volume =. Journal of Multivariate Analysis , author =. 2002 , keywords =

    2002

  20. [28]

    2024 , author =

    Nearest neighbors weighted composite likelihood based on pairs for (non-)Gaussian massive spatial data with an application to Tukey-hh random fields estimation , journal =. 2024 , author =

    2024

  21. [29]

    Unifying compactly supported and Mat

    Bevilacqua, Moreno and Caama. Unifying compactly supported and Mat. Journal of Multivariate Analysis , volume=. 2022 , publisher=

    2022

  22. [30]

    Ma, Pulong and Bhadra, Anindya , journal=. Beyond. 2022 , publisher=

    2022

  23. [31]

    and Delfiner, P

    Chil \` e s, J.P. and Delfiner, P. Geostatistics: M odeling S patial U ncertainty. 2012

    2012

  24. [32]

    Keyes and Marc G

    Sameh Abdulah and Faten Alamri and Pratik Nag and Ying Sun and Hatem Ltaief and David E. Keyes and Marc G. Genton , title =. Journal of Data Science , volume =. 2022 , pages =

    2022

  25. [33]

    M. J. Heaton and A. Datta and A. O. Finley and R. Furrer and J. Guinness and R. Guhaniyogi and F. Gerber and R. B. Gramacy and D. Hammerling and M. Katzfuss and F. Lindgren and D. W. Nychka and F. Sun and A. Zammit-Mangion , title =. Journal of Agricultural, Biological, and En...

  26. [34]

    Huang and S

    H. Huang and S. Abdulah and Y. Sun and L. Hatem and D. Keyes and M. Genton , title =. Journal of Agricultural, Biological, and Environmental Statistics , year =

  27. [35]

    Journal of the American Statistical Association , volume =

    Tilmann Gneiting and William Kleiber and Martin Schlather , title =. Journal of the American Statistical Association , volume =. 2010 , publisher =

    2010

  28. [36]

    Nonseparable stationary anisotropic space--time covariance functions , Volume =

    Emilio Porcu and Pablo Gregori and Jorge Mateu , Date-Added =. Nonseparable stationary anisotropic space--time covariance functions , Volume =. Stoch. Environ. Res. Risk. Assess. , Pages =

  29. [37]

    Genton , Date-Added =

    Huiyan Sang and Marc G. Genton , Date-Added =. Tapered Composite Likelihood for Spatial Max-Stable Models , Year =

  30. [38]

    Fuentes , Date-Added =

    M. Fuentes , Date-Added =. Approximate likelihood for large irregularly spaced spatial data , Volume =. Journal of the American Statistical Association , Pages =

  31. [39]

    Minka , Date-Added =

    Ray Brownrigg and Thomas P. Minka , Date-Added =. 2013 , Bdsk-Url-1 =

    2013

  32. [40]

    Banerjee , Date-Added =

    S. Banerjee , Date-Added =. On Geodetic Distance Computations in Spatial Modeling , Volume =. Biometrics , Pages =

  33. [41]

    Simulation and Analysis of Random Fields , Volume =

    Martin Schlather , Date-Added =. Simulation and Analysis of Random Fields , Volume =

  34. [42]

    2013 , Bdsk-Url-1 =

    Martin Schlather , Date-Added =. 2013 , Bdsk-Url-1 =

    2013

  35. [43]

    and Zeger, S.L

    Heagerty, P. and Zeger, S.L. , Date-Added =. Lorelogram: A regression Approach to Exploring Association in Longitudinal Categorical Responses , Volume =. Journal of the American Statistical Association, Theory & Methods , Number =

  36. [44]

    and Ruppert, D

    Shaby, B. and Ruppert, D. , Date-Added =. Tapered covariance: Bayesian Estimation and Asymptotics , Volume =. Journal of Computational and Graphical Statistics , Pages =

  37. [45]

    Sain , Date-Added =

    Reinhard Furrer and Stephan R. Sain , Date-Added =. Journal of Statistical Software , Pages =

  38. [46]

    2015 , pages=

    Journal of Statistical Software , author=. 2015 , pages=

    2015

  39. [47]

    2022 , pages=

    Journal of Statistical Software , author=. 2022 , pages=

    2022

  40. [48]

    2021 , pages=

    Journal of Statistical Software , author=. 2021 , pages=

    2021

  41. [49]

    Statistics and Computing , year =

    James, Zachary and Guinness, Joseph , title =. Statistics and Computing , year =. doi:10.1007/s11222-024-10510-9 , url =

    doi

  42. [50]

    Guinness, Joseph , year =

  43. [51]

    Modeling Big, Heterogeneous, Non-Gaussian Spatial and Spatio-Temporal Data Using

    Sainsbury-Dale Matthew and Andrew Zammit-Mangion and Noel Cressie , Journal =. Modeling Big, Heterogeneous, Non-Gaussian Spatial and Spatio-Temporal Data Using

  44. [52]

    2013 , Bdsk-Url-1 =

    Reinhard Furrer , Date-Added =. 2013 , Bdsk-Url-1 =

    2013

  45. [53]

    Poncet , Date-Added =

    Daniel Cooley and Philippe Neveau and P. Poncet , Date-Added =. Variograms for Spatial Max-stable Random Fields , Volume =

  46. [54]

    The Annals of Statistics , number =

    Moreno Bevilacqua and Tarik Faouzi and Reinhard Furrer and Emilio Porcu , title =. The Annals of Statistics , number =

  47. [55]

    Wendland , Date-Added =

    H. Wendland , Date-Added =. Piecewise Polynomial, Positive Definite and Compactly Supported Radial Functions of Minimal Degree , Volume =. Advances in Computational Mathematics , Pages =

  48. [56]

    K. V. Mardia and R. J. Marshall , Date-Added =. Maximum Likelihood Estimation of Models for Residual Covariance in Spatial Regression , Volume =. Biometrika , Number =

  49. [57]

    D. A. Harville , Date-Added =. Maximum Likelihood Approaches to Variance Component Estimation and to Related Problems , Volume =. Journal of the American Statistical Association, Theory & Methods , Pages =

  50. [58]

    Adjusting Composite Likelihood Ratio Statistic , Volume =

    Luigi Pace and Alessandra Salvan and Nicola Sartori , Date-Added =. Adjusting Composite Likelihood Ratio Statistic , Volume =. Statistica Sinica , Pages =

  51. [59]

    Davis and Claudia Kl\"

    Richard A. Davis and Claudia Kl\". Journal of the Korean Statistical Society, http://dx.doi.org/10.1016/j.bbr.2011.03.031 , Title =

    2011 doi

  52. [60]

    Opitz , Date-Added =

    T. Opitz , Date-Added =. Extremal t processes: Elliptical domain of attraction and a spectral representation , Volume =. Journal of Multivariate Analysis , Pages =

  53. [61]

    Stein , Date-Added =

    Micheal L. Stein , Date-Added =. Space-Time Covariance Functions , Volume =. Journal of the American Statistical Association, Theory & Methods , Number =

  54. [62]

    Quasi-Arithmetic Means of Covariance Functions with Potential Applications to Space-Time Data , Volume =

    Emilio Porcu and Jorge Mateu and George Christakos , Date-Added =. Quasi-Arithmetic Means of Covariance Functions with Potential Applications to Space-Time Data , Volume =. Journal of Multivariate Analysis , Number =

  55. [63]

    and Genton ,M.G

    Gneiting,T. and Genton ,M.G. and Guttorp, P. , Date-Added =. Geostatistical space-time models, stationarity, separability and full symmetry. In Statistical Methods for Spatio-Temporal Systems , Year =

  56. [64]

    Spatial Data Science: With Applications in

    Edzer Pebesma and Roger Bivand , Publisher =. Spatial Data Science: With Applications in

  57. [65]

    Journal of Faculty of Civil Engineering , Title =

    Mirko Borisov and Milan Vrtunski and Vladimir Petrovi. Journal of Faculty of Civil Engineering , Title =

  58. [66]

    Nonseparable, stationary covariance functions for space-time data , Volume =

    Tilman Gneiting , Date-Added =. Nonseparable, stationary covariance functions for space-time data , Volume =. Journal of the American Statistical Association , Number =

  59. [67]

    and Keyes, David E

    Abdulah, Sameh and Ltaief, Hatem and Sun, Ying and Genton, Marc G. and Keyes, David E. , journal=. ExaGeoStat: A High Performance Unified Software for Geostatistics on Manycore Systems , year=

  60. [68]

    Diggle , Date-Added =

    Alexandre Rodrigues and Peter J. Diggle , Date-Added =. A Class of Convolution-based Models for Spatio-temporal Processes with Non-separable Covariance Structure , Volume =. Scandinavian Journal of Statistics , Pages =

  61. [69]

    Bivand , Journal =

    R.S. Bivand , Journal =. Progress in the

  62. [70]

    Andreas Papritz , year =

  63. [71]

    2022 , note =

    Paulo Justiniano. 2022 , note =

    2022

  64. [72]

    Statistica Sinica , volume=

    A class of multi-resolution approximations for large spatial datasets , author=. Statistica Sinica , volume=. 2020 , publisher=

    2020

  65. [73]

    Furrer and R

    R. Furrer and R. Flury and F. Gerber , year =

  66. [74]

    Journal of the American Statistical Association , volume =

    Matthias Katzfuss , title =. Journal of the American Statistical Association , volume =

  67. [75]

    Technometrics , volume =

    Joseph Guinness , title =. Technometrics , volume =. 2018 , publisher =

    2018

  68. [76]

    , author =

    CRAN Task View: Analysis of Spatial Data. , author =. 2024 , note =

    2024

  69. [77]

    Katzfuss and M

    M. Katzfuss and M. Jurek and D. Zilber and W. Gong and J. Guinness and J. Zhang and F. Sch. 2020 , note =

    2020

  70. [78]

    2016 , url =

    Douglas Nychka and Dorit Hammerling and Stephan Sain and Nathan Lenssen , note =. 2016 , url =

    2016

  71. [79]

    2008 , journal =

    Fixed rank kriging for very large spatial data sets , author =. 2008 , journal =

    2008

  72. [80]

    Spatio-Temporal Interpolation using

    Benedikt Gr. Spatio-Temporal Interpolation using. 2016 , journal =

    2016

  73. [81]

    2008 , journal =

    Gaussian predictive process models for large spatial datasets , author =. 2008 , journal =

    2008

  74. [82]

    2009 , author =

    Improving the performance of predictive process modeling for large datasets , journal =. 2009 , author =

    2009

  75. [83]

    Keyes , booktitle=

    Sameh Abdulah and Hatem Ltaief and Ying Sun and Marc Genton and David E. Keyes , booktitle=. Parallel Approximation of the Maximum Likelihood Estimation for the Prediction of Large-Scale Geostatistics Simulations , year=

  76. [84]

    , title =

    Mantoglou, Aristotelis and Wilson, John L. , title =. Water Resources Research , volume =

  77. [85]

    Blanchet , Date-Added =

    Anthony Christian Davison and J. Blanchet , Date-Added =. Spatial modelling of extreme snow depth , Volume =. Annals of Applied Statistics , Number =

  78. [86]

    Environmetrics , author =

    Composite likelihood estimation for models of spatial ordinal data and spatial proportional data with zero/one values , volume =. Environmetrics , author =. 2014 , pages =

    2014

  79. [87]

    Gradshteyn and I.M

    I.S. Gradshteyn and I.M. Ryzhik , Publisher =. Table of Integrals, Series, and Products , Year =

  80. [88]

    Journal of Software Engineering for Robotics (JOSER) , pages =

    Comparison of nearest-neighbor-search strategies and implementations for efficient shape registration , author =. Journal of Software Engineering for Robotics (JOSER) , pages =. 2012 , issn =

    2012

  81. [89]

    J. L. Bentley , title =. 1975 , volume =

    1975

  82. [90]

    Arya and D

    S. Arya and D. M. Mount and N. S. Netanyahu and R. Silverman and A. Y. Wu , title =. 1998 , volume =

    1998

  83. [91]

    Zilber and M

    D. Zilber and M. Katzfuss , title =. 2021 , volume =

    2021

  84. [92]

    Davis and Chun Yip Yau , journal =

    Richard A. Davis and Chun Yip Yau , journal =. Comments on pairwise likelihood in time series models , urldate =

  85. [93]

    Vecchia, A. V. , title = ". Journal of the Royal Statistical Society: Series B (Methodological) , volume =. 1988 , month =

    1988

  86. [94]

    Statistical Science , number =

    Matthias Katzfuss and Joseph Guinness , title =. Statistical Science , number =

  87. [95]

    arXiv preprint arXiv:2008.00553 , year=

    A unifying framework for parallel and distributed processing in R using futures , author=. arXiv preprint arXiv:2008.00553 , year=

    2008 arXiv

  88. [96]

    , journal=

    Stein, Michael L. , journal=. 2010. 2011 , doi=

    2010

  89. [97]

    arXiv preprint arXiv:2512.18884 , year=

    Fast simulation of Gaussian random fields with flexible correlation models in Euclidean spaces , author=. arXiv preprint arXiv:2512.18884 , year=

  90. [98]

    Genton and William Kleiber , title =

    Marc G. Genton and William Kleiber , title =. Statistical Science , number =. 2015 , doi =

    2015

  91. [99]

    WIREs Computational Statistics , volume =

    Porcu, Emilio and Furrer, Reinhard and Nychka, Douglas , title =. WIREs Computational Statistics , volume =

  92. [100]

    2026 , note =

    Bevilacqua, Moreno and Morales-O. 2026 , note =

    2026

  93. [101]

    Comparing composite likelihood methods based on pairs for spatial Gaussian random fields , keywords=

    Bevilacqua, Moreno and Gaetan, Carlo , volume =. Comparing composite likelihood methods based on pairs for spatial Gaussian random fields , keywords=. 2015 , journal=

    2015

  94. [102]

    A class of random fields with two-piece marginal distributions for modeling point-referenced data with spatial outliers , journal =

    Moreno Bevilacqua and Christian Caama. A class of random fields with two-piece marginal distributions for modeling point-referenced data with spatial outliers , journal =. 2022 , volume =

    2022

  95. [103]

    A selective view of climatological data and likelihood estimation , journal =

    Federico Blasi and Christian Caama. A selective view of climatological data and likelihood estimation , journal =. 2022 , issn =

    2022

  96. [104]

    Bevilacqua and C

    M. Bevilacqua and C. Caama. Non-Gaussian geostatistical modeling using (skew) t processes , journal =

  97. [105]

    Jones, M. C. and Pewsey, Arthur , title = ". Biometrika , volume =. 2009 , month =

    2009

  98. [106]

    and Jeong, J

    Yan, Y. and Jeong, J. and Genton, M.G. , journal =. Multivariate transformed Gaussian processes , year =

  99. [107]

    Azzalini , journal =

    A. Azzalini , journal =. A Class of Distributions Which Includes the Normal Ones , urldate =

  100. [108]

    Environmetrics: The official journal of the International Environmetrics Society , volume=

    Kriging and cross-validation for massive spatial data , author=. Environmetrics: The official journal of the International Environmetrics Society , volume=. 2010 , publisher=

    2010

  101. [109]

    2017 , author =

    Estimating covariance functions of multivariate skew-Gaussian random fields on the sphere , journal =. 2017 , author =

    2017

  102. [110]

    Environmetrics , volume =

    Zhang, Hao and El-Shaarawi, Abdel , title =. Environmetrics , volume =

  103. [111]

    On modeling positive continuous data with spatiotemporal dependence , journal =

    Moreno Bevilacqua and Christian Caama. On modeling positive continuous data with spatiotemporal dependence , journal =

  104. [112]

    Spatial Statistics , volume=

    Poisson kriging: A closer investigation , author=. Spatial Statistics , volume=. 2014 , publisher=

    2014

  105. [113]

    GeoENV VII--geostatistics for environmental applications , pages=

    Geostatistical modelling of wildlife populations: a non-stationary hierarchical model for count data , author=. GeoENV VII--geostatistics for environmental applications , pages=. 2010 , publisher=

    2010

  106. [114]

    Bayesian Prediction of Transformed Gaussian Random Fields , urldate =

    V. Bayesian Prediction of Transformed Gaussian Random Fields , urldate =. Journal of the American Statistical Association , number =

  107. [115]

    Allcroft and C.A

    D.J. Allcroft and C.A. Glasbey , title =. Journal of the Royal Statistical Society, C , year =

  108. [116]

    Scandinavian Journal of Statistics , author =

    On optimal point and block prediction in log-gaussian random fields , volume =. Scandinavian Journal of Statistics , author =. 2006 , pages =

    2006

  109. [117]

    Lele and M

    Subhash R. Lele and M. L. Taper , Date-Added =. A Composite Likelihood Approach to (co)variance Components Estimation , Volume =. Journal of Statistical Planning and Inference , Pages =

  110. [118]

    Furrer and Marc G

    R. Furrer and Marc G. Genton and D. Nychka , Date-Added =. Covariance Tapering for Interpolation of Large Spatial Datasets , Volume =. Journal of Computational and Graphical Statistics , Pages =

  111. [119]

    Xavier Emery and Alfredo Alegr. The. 2022 , journal=

    2022

  112. [120]

    Spatial Statistics , volume =

    The F-family of covariance functions: A Mat. Spatial Statistics , volume =. 2021 , author =

    2021

  113. [121]

    Journal of the American Statistical Association , volume =

    Pulong Ma and Anindya Bhadra , title =. Journal of the American Statistical Association , volume =. 2022 , publisher =

    2022

  114. [122]

    Statistical Science , author =

    The Mat. Statistical Science , author =

  115. [123]

    Some Theory of Kriging , publisher =

    Interpolation of Spatial Data. Some Theory of Kriging , publisher =. 1999 , author =

    1999

  116. [124]

    Stein and Z

    Micheal L. Stein and Z. Chi and L. J. Welty , Date-Added =. Approximating likelihoods for large spatial data sets , Volume =. Journal of the Royal Statistical Society, Series B , Pages =

  117. [125]

    Johannesson , Date-Added =

    Nohal Cressie and G. Johannesson , Date-Added =. Fixed Rank Kriging for Very Large Spatial Data Sets , Volume =. Journal of the Royal Statistical Society, Series B , Pages =

  118. [126]

    1970 , editor =

    Handbook of Mathematical Functions , publisher =. 1970 , editor =

    1970

  119. [127]

    D. L. Zimmerman , Date-Added =. Computationally Exploitable Structure of Covariance Matrices and Generalized Covariance Matrices in Spatial Models , Volume =. Journal of Statistical Computation and Simulation , Pages =

  120. [128]

    Diggle and Paulo J

    Peter J. Diggle and Paulo J. Ribeiro , Date-Added =. Model-based Geostatistics , Year =

  121. [129]

    Vallejos and F

    R. Vallejos and F. Osorio and M. Bevilacqua , Publisher =. Spatial Relationships between Two Georeferenced Variables: With Applications in R , Year =

  122. [130]

    Doug McIlroy , year =

  123. [131]

    Nott, D. J. and Ryd\'en, T. , Date-Added =. Pairwise Likelihood Methods for Inference in Image Models , Volume =. Biometrika , Pages =

  124. [132]

    Kaufman, C. G. and Schervish, M. J. and Nychka, D. W. , Journal =. Covariance Tapering for Likelihood--Based Estimation in Large Spatial Data Sets , Volume =

  125. [133]

    and Shaby, Benjamin A

    Kaufman, Cari G. and Shaby, Benjamin A. , title =. Biometrika , year =

  126. [134]

    and Reid, N

    Varin, C. and Reid, N. and Firth, D. , Date-Added =. An Overview of Composite Likelihood Methods , Volume =. Statistica Sinica , Number =

  127. [135]

    2021 , issn =

    Vecchia--Laplace approximations of generalized Gaussian processes for big non-Gaussian spatial data , journal =. 2021 , issn =

    2021

  128. [136]

    Diggle and J.A

    P.J. Diggle and J.A. Tawn and R.A. Moyeed , title =. Journal of the Royal Statistical Society: Series C. 1998 , volume =

    1998

  129. [137]

    The Annals of Applied Statistics , number =

    Annamaria Guolo and Cristiano Varin , title =. The Annals of Applied Statistics , number =

  130. [138]

    Stochastic Environmental Research and Risk Assessment , author =

    Copula-based geostatistical modeling of continuous and discrete data including covariates , volume =. Stochastic Environmental Research and Risk Assessment , author =. 2010 , pages =

    2010

  131. [139]

    Generalized Additive Models: An Introduction with

    Wood, Simon , year =. Generalized Additive Models: An Introduction with

  132. [140]

    Journal of Statistical Software , author=

    Bayesian Spatial Modelling with. Journal of Statistical Software , author=. 2015 , pages=

    2015

  133. [141]

    An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach , journal =

    Lindgren, Finn and Rue, H. An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach , journal =

  134. [142]

    2016 , author =

    Spatial statistics and Gaussian processes: A beautiful marriage , journal =. 2016 , author =

    2016

  135. [143]

    Scandinavian Journal of Statistics , author =

    Geostatistical Modelling using Non-Gaussian Mat\'ern Fields , volume =. Scandinavian Journal of Statistics , author =. 2015 , pages =

    2015

  136. [144]

    Davison, A. C. and Gholamrezaee, M. M. , Date-Added =. Geostatistics of extremes , Volume =. Proceedings of the Royal Society of London, series A , Number =

  137. [145]

    , Date-Added =

    Kabluchko, Z. , Date-Added =. Extremes of Space-Time Gaussian Processes , Volume =. Stochastic Processes and their Applications , Pages =

  138. [146]

    2017 , journal =

    Tukey g- and -h Random Fields , author =. 2017 , journal =

    2017

  139. [147]

    Song , title =

    Xin Gao and Peter X.-K. Song , title =. Journal of the American Statistical Association , volume =. 2010 , publisher =

    2010

  140. [148]

    Journal of the American Statistical Association , author =

    A Composite Likelihood Approach to Binary Spatial Data , volume =. Journal of the American Statistical Association , author =. 1998 , pages =

    1998

  141. [149]

    Electronic Journal of Statistics , number =

    Moreno Bevilacqua and Xavier Emery and Tarik Faouzi , title =. Electronic Journal of Statistics , number =

  142. [150]

    de Leon, A. R. , Date-Added =. Pairwise likelihood approach to grouped continuous model and its extension , Volume =. Statistics and Probability Letters , Number =

  143. [151]

    Multivariate Geostatistics: An Introduction with Applications , Year =

    Henric Wackernagel , Date-Added =. Multivariate Geostatistics: An Introduction with Applications , Year =

  144. [152]

    Random Field Models in Earth Sciences , Year =

    George Christakos , Date-Added =. Random Field Models in Earth Sciences , Year =

  145. [153]

    Finley and Alan E

    Abhirup Datta, Sudipto Banerjee, Andrew O. Finley and Alan E. Gelfand , title =. Journal of the American Statistical Association , volume =. 2016 , publisher =

    2016

  146. [154]

    Carlin and Alan E

    Sudipto Banerjee and Bradley P. Carlin and Alan E. Gelfand , Publisher =. Hierarchical Modeling and Analysis for Spatial Data 2nd Edition , Year =

  147. [155]

    Padoan and Mathieu Ribatet , Date-Added =

    Anthony Christian Davison and Simone A. Padoan and Mathieu Ribatet , Date-Added =. Statistical Modelling of Spatial Extremes , Volume =. Statistical Science , Number =

  148. [156]

    Genton, M. G. and Ma, Y. and Sang, H. , Date-Added =. On the Likelihood Function of Gaussian Max-Stable Processes Indexed by R^d , d 1 , Volume =. Biometrika , Number =

  149. [157]

    Davis and Philippe Neveau , Date-Added =

    Daniel Cooley and Richard A. Davis and Philippe Neveau , Date-Added =. The Pairwise Beta Distribution: A Flexible Parametric Multivariate Model for Extremes , Volume =. Journal of Multivariate Analysis , Pages =

  150. [158]

    Padoan and Mathieu Ribatet and Scott A

    Simone A. Padoan and Mathieu Ribatet and Scott A. Sisson , Date-Added =. Likelihood-Based Inference for Max-Stable Processes , Volume =. Journal of the American Statistical Association, Theory & Methods , Pages =

  151. [159]

    Estimating Space and Space--Time Covariance Functions: a Weighted Composite Likelihood Approach , Volume =

    Moreno Bevilacqua and Carlo Gaetan and Jotge Mateu and Emilio Porcu , Date-Added =. Estimating Space and Space--Time Covariance Functions: a Weighted Composite Likelihood Approach , Volume =. Journal of the American Statistical Association, Theory & Methods , Number =

  152. [160]

    B. M. Brown and S. I. Resnick , Date-Added =. Extreme Values of Independent Stochastic Processes , Volume =. Journal of Applied Probability , Pages =

  153. [161]

    Stationary Max-Stable Fields Associated to Negative Definite Functions , Volume =

    Zakhar Kabluchko and Martin Schlather and Laurent. Stationary Max-Stable Fields Associated to Negative Definite Functions , Volume =. The Annals of Probability , Number =

  154. [162]

    Extremes of Independent Gaussian Processes , Volume =

    Zakar Kabluchko , Date-Added =. Extremes of Independent Gaussian Processes , Volume =. Extremes , Pages =

  155. [163]

    J. H\". Maxima of Normal Random Vectors: Between Independence and Complete Dependence , Volume =. Statistics and Probability Letters , Pages =

  156. [164]

    J. H\". Maxima of normal random vectors: between independence and complete dependence , Volume =. Statistics and Probability Letters , Pages =

  157. [165]

    Heyde , Date-Added =

    C. Heyde , Date-Added =. Quasi-Likelihood and Its Application: A General Approach to Optimal Parameter Estimation , Year =

  158. [166]

    Cressie and Christopher K

    Noel A. Cressie and Christopher K. Wikle , Date-Added =. Statistics for Spatial-Temporal Data , Year =

  159. [167]

    S. G. Coles , Date-Added =. An Introduction to Statistical Modelling of Extreme Values , Year =

  160. [168]

    Takeuchi , Date-Added =

    K. Takeuchi , Date-Added =. Distribution of information statistics and criteria for adequacy of models (in Japenese) , Volume =. Mathematical Science , Pages =

  161. [169]

    R. E. Chandler and S. Bate , Date-Added =. Inference for Clustered Data Using the Independence Loglikelihood , Volume =. Biometrika , Pages =

  162. [170]

    Rotnitzky and N

    A. Rotnitzky and N. P. Jewell , Date-Added =. Hypothesis-testing of regression parameters in semiparametric generalised linear models for cluster correlated data. , Volume =. Biometrika , Pages =

  163. [171]

    2024 , author =

    A flexible Clayton-like spatial copula with application to bounded support data , journal =. 2024 , author =

    2024

  164. [172]

    Journal of Statistical Software , author=

    Gaussian Copula Regression in. Journal of Statistical Software , author=. 2017 , pages=

    2017

  165. [173]

    Water Resources Research , volume=

    Copula-based geostatistical models for groundwater quality parameters , author=. Water Resources Research , volume=. 2006 , publisher=

    2006

  166. [174]

    J. T. Kent , Date-Added =. Information gain and a measure of correlation , Volume =. Biometrika , Pages =

  167. [175]

    Varin and P

    C. Varin and P. Vidoni , title =. Biometrika , year =

  168. [176]

    V. P. Godambe , Date-Added =. An optimum property of regular maximum likelihood estimation , Volume =. Annals of Mathematical Statistics , Pages =

  169. [177]

    D. R. Cox and N. Reid , Date-Added =. A Note on Pseudolikelihood Constructed from Marginal Densities , Volume =. Biometrika , Pages =

  170. [178]

    Journal of the American Statistical Association , volume =

    Gildas Mazo and Dimitris Karlis and Andrea Rau , title =. Journal of the American Statistical Association , volume =. 2024 , publisher =

    2024

  171. [179]

    Journal of the American Statistical Association , volume =

    Giuseppe Alfonzetti and Ruggero Bellio and Yunxiao Chen and Irini Moustaki , title =. Journal of the American Statistical Association , volume =. 2025 , publisher =

    2025

  172. [180]

    B. G. Lindsay , Date-Added =. Composite Likelihood Methods , Volume =. Contemporary Mathematics , Pages =

  173. [181]

    Varin , Date-Added =

    C. Varin , Date-Added =. On composite Marginal Likelihoods , Volume =. Advances in Statistical Analysis , Pages =

  174. [182]

    Spatial extremes:

    Laurent. Spatial extremes:. The Annals of Statistics , Pages =

  175. [183]

    Extreme Value Theory An Introduction , Year =

    Laurent. Extreme Value Theory An Introduction , Year =

  176. [184]

    Dependence modeling with copulas , Year =

    Harry Joe , Publisher =. Dependence modeling with copulas , Year =

  177. [185]

    2007 , author =

    Model-based Geostatistics , publisher =. 2007 , author =

    2007

  178. [186]

    Model-based Geostatistics for Global Public Health: Methods and Applications , isbn =

    Diggle, Peter and Giorgi, Emanuele , year =. Model-based Geostatistics for Global Public Health: Methods and Applications , isbn =

  179. [187]

    Morales-Navarrete and M

    D. Morales-Navarrete and M. Bevilacqua and C. Caama. Modeling Point Referenced Spatial Count Data: A Poisson Process Approach , journal =. 2022 , publisher =

    2022

  180. [188]

    Resnick , Date-Added =

    S. Resnick , Date-Added =. Extreme Values, Point Processes and Regular Variation , Year =

  181. [189]

    , Date-Added =

    Schlather, M. , Date-Added =. Models for Stationary Max-Stable Random Fields , Volume =. Extremes , Number =

  182. [190]

    Tawn , Date-Added =

    Martin Schlather and Jonathan A. Tawn , Date-Added =. A dependence measure for multivariate and spatial extreme values:. Biometrika , Pages =

  183. [191]

    R. L. Smith , Date-Added =. Max-Stable Processes and Spatial Extremes , Year =

  184. [192]

    Stationary min-stable stochastic processes , Volume =

    Laurent. Stationary min-stable stochastic processes , Volume =. Probability Theory and Related Fields , Pages =

  185. [193]

    R. A. Fisher and L. H. C. Tippett , Date-Added =. Limiting forms of the frequency distribution of the largest or smallest member of a sample , Volume =. Proceedings of the Cambridge Philosophical Society , Pages =

  186. [194]

    Limit theory for multivariate sample extremes , Volume =

    Laurent. Limit theory for multivariate sample extremes , Volume =. Z. Wahrscheinlichkeitstheorie verw. Gebiete , Pages =

  187. [195]

    P. F. Verdes , Date-Added =. Global warming is driven by anthropogenic emissions:. Physical Review Letters , Number =

  188. [196]

    A spectral representation for max-stable processes , Volume =

    Laurent. A spectral representation for max-stable processes , Volume =. The Annals of Probability , Pages =

  189. [197]

    International Journal for Uncertainty Quantification , volume=

    A block circulant embedding method for simulation of stationary Gaussian random fields on block-regular grids , author=. International Journal for Uncertainty Quantification , volume=. 2015 , publisher=

    2015

Pith tools

Reviewed July 8, 2026 · model on record in the stance chip above.