pith. sign in

arxiv: 2605.23760 · v1 · pith:4SHSUCR7new · submitted 2026-05-22 · 📊 stat.ME

Global Sensitivity Analysis: a novel generation of mighty estimators based on rank statistics

Fabrice Gamboa (IMT) , Pierre Gremaud (NC State) , Thierry Klein (IMT , ENAC) , Agn\`es Lagnoux (IMT) This is my paper

Pith reviewed 2026-05-25 03:19 UTC · model grok-4.3

classification 📊 stat.ME
keywords global sensitivity analysisSobol indicesrank statisticsChatterjee correlationconsistencycentral limit theoremestimation
0
0 comments X

The pith

Rank statistics using Chatterjee correlation yield consistent estimators for Sobol indices and related sensitivity measures

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

The paper develops an estimation framework for global sensitivity analysis indices that applies rank statistics through an empirical version of Chatterjee's correlation coefficient. The same approach produces estimators for Cramér-von-Mises indices, first-order Sobol indices, indices defined on general metric spaces, and higher-order moment indices. Consistency is proved for all these estimators, numerical performance is demonstrated especially at small sample sizes, and a central limit theorem is established specifically for the first-order Sobol estimators. A reader would care because the method supplies a single, rank-based procedure that covers several important indices without requiring large samples or model-specific tuning.

Core claim

The authors show that a large family of global sensitivity indices can be estimated consistently by replacing the usual integrals or expectations with an empirical correlation coefficient built from rank statistics, as introduced by Chatterjee. This produces estimators for Cramér-von-Mises indices, first-order Sobol indices, general metric-space indices and higher-order moment indices. The estimators are consistent, numerically efficient for small samples, and the Sobol-index versions satisfy a central limit theorem.

What carries the argument

Chatterjee's empirical correlation coefficient computed from rank statistics, which directly supplies the estimators for the listed sensitivity indices

If this is right

  • The estimators remain consistent under standard i.i.d. sampling
  • Numerical performance is strong even with small sample sizes
  • The first-order Sobol estimators obey a central limit theorem
  • The same rank-statistic construction covers metric-space and higher-order moment indices

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method may lower the data requirement for sensitivity analysis in expensive simulation models
  • Because ranks are invariant to monotonic transformations, the estimators could remain stable under reparameterizations of the inputs
  • Direct comparison on benchmark functions with known indices would quantify the small-sample gain relative to Monte Carlo or polynomial-chaos alternatives

Load-bearing premise

The target sensitivity indices can be expressed exactly in terms of Chatterjee's rank-based correlation coefficient

What would settle it

In a low-dimensional test function whose true first-order Sobol indices are known analytically, the new estimator fails to converge to those values as the number of independent samples tends to infinity

read the original abstract

We propose a new statistical estimation framework for a large family of global sensitivity analysis indices. Our approach is based on rank statistics and uses an empirical correlation coefficient recently introduced by Chatterjee [9]. We show how to apply this approach to compute not only the Cram{\'e}r-von-Mises indices, which are directly related to Chatterjee's notion of correlation, but also first-order Sobol indices, general metric space indices and higher-order moment indices. We establish consistency of the resulting estimators and demonstrate their numerical efficiency, especially for small sample sizes. In addition, we prove a central limit theorem for the estimators of the first-order Sobol indices.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper proposes a new statistical estimation framework for global sensitivity analysis indices based on rank statistics and Chatterjee's empirical correlation coefficient. It shows applications to Cramér-von-Mises indices, first-order Sobol indices, general metric-space indices, and higher-order moment indices. The authors establish consistency of the resulting estimators, demonstrate numerical efficiency for small sample sizes, and prove a central limit theorem for the first-order Sobol index estimators.

Significance. If the claimed consistency, CLT, and efficiency results hold, the work supplies a unified rank-based approach to estimating multiple families of sensitivity indices with explicit constructions and theoretical support. This is potentially valuable for applications requiring reliable estimates from limited data, and the machine-checked or explicitly derived proofs (as indicated in the full manuscript) add to its strength.

minor comments (3)
  1. [§3.2] §3.2: the notation for the empirical Chatterjee coefficient is introduced without an explicit comparison table to the population version; adding this would clarify the convergence argument.
  2. [Figure 2] Figure 2: the caption does not state the sample size or number of replications used in the small-sample comparison; this information is needed to interpret the efficiency claims.
  3. The reference list omits the original Chatterjee (2021) paper on the rank correlation; it should be cited directly rather than only via secondary sources.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive summary of the manuscript and for recommending minor revision. No specific major comments were listed in the report, so we have no individual points requiring point-by-point response or revision.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The derivation applies Chatterjee's external rank-based correlation (citation [9]) to construct estimators for Cramér-von-Mises, Sobol, metric-space, and moment indices, then proves consistency and a CLT under standard i.i.d. assumptions. No step reduces by definition to a fitted parameter, self-citation chain, or ansatz smuggled from prior author work; the mapping from rank statistics to each index is explicitly constructed and externally benchmarked rather than tautological. The central claims remain independent of the paper's own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides insufficient detail to enumerate specific free parameters, axioms, or invented entities; the work relies on standard statistical sampling assumptions and the properties of Chatterjee's coefficient.

pith-pipeline@v0.9.0 · 5659 in / 1047 out tokens · 32477 ms · 2026-05-25T03:19:02.182301+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages · 1 internal anchor

  1. [1]

    Antoniadis

    A. Antoniadis. Analysis of variance on function spaces. Statistics: A Journal of Theoretical and Applied Statistics , 15(1):59--71, 1984

    work page 1984

  2. [2]

    Auddy, N

    A. Auddy, N. Deb, and S. Nandy. Exact detection thresholds for chatterjee's correlation. arXiv preprint arXiv:2104.15140 , 2021

    work page arXiv 2021

  3. [3]

    Azadkia and S

    M. Azadkia and S. Chatterjee. A simple measure of conditional dependence. The Annals of Statistics , 49(6):3070--3102, 2021

    work page 2021

  4. [4]

    Borgonovo

    E. Borgonovo. A new uncertainty importance measure. Reliability Engineering & System Safety , 92(6):771--784, 2007

    work page 2007

  5. [5]

    Borgonovo, W

    E. Borgonovo, W. Castaings, and S. Tarantola. Moment independent importance measures: New results and analytical test cases. Risk Analysis , 31(3):404--428, 2011

    work page 2011

  6. [6]

    Boucheron, G

    S. Boucheron, G. Lugosi, and P. Massart. Concentration inequalities: A nonasymptotic theory of independence . Oxford university press, 2013

    work page 2013

  7. [7]

    Boucheron, G

    S. Boucheron, G. Lugosi, P. Massart, et al. On concentration of self-bounding functions. Electronic Journal of Probability , 14:1884--1899, 2009

    work page 2009

  8. [8]

    Broto, F

    B. Broto, F. Bachoc, and M. Depecker. Variance reduction for estimation of shapley effects and adaptation to unknown input distribution. SIAM/ASA Journal on Uncertainty Quantification , 8(2):693--716, 2020

    work page 2020

  9. [9]

    Chatterjee

    S. Chatterjee. A new coefficient of correlation. Journal of the American Statistical Association , pages 1--26, 2020

    work page 2020

  10. [10]

    Da Veiga

    S. Da Veiga. Global sensitivity analysis with dependence measures. J. Stat. Comput. Simul. , 85(7):1283--1305, 2015

    work page 2015

  11. [11]

    Da Veiga and F

    S. Da Veiga and F. Gamboa. Efficient estimation of sensitivity indices. Journal of Nonparametric Statistics , 25(3):573--595, 2013

    work page 2013

  12. [12]

    Da Veiga, F

    S. Da Veiga, F. Gamboa, A. Lagnoux, T. Klein, and C. Prieur. Efficient estimation of Sobol' indices of any order froma single input/output sample . Submitted paper, 2025

    work page 2025

  13. [13]

    De Rocquigny, N

    E. De Rocquigny, N. Devictor, and S. Tarantola. Uncertainty in industrial practice . Wiley Online Library, 2008

    work page 2008

  14. [14]

    Dette, K

    H. Dette, K. F. Siburg, and P. A. Stoimenov. A copula-based non-parametric measure of regression dependence. Scandinavian Journal of Statistics , 40(1):21--41, 2013

    work page 2013

  15. [15]

    J.-C. Fort, T. Klein, and A. Lagnoux. Global sensitivity analysis and wasserstein spaces. SIAM/ASA Journal on Uncertainty Quantification , 9(2):880--921, 2021

    work page 2021

  16. [16]

    Fort , T

    J.-C. Fort , T. Klein , and N. Rachdi . New sensitivity analysis subordinated to a contrast . ArXiv e-prints , May 2013

    work page 2013

  17. [17]

    Sensitivity indices for output on a Riemannian manifold

    R. Fraiman , F. Gamboa , and L. Moreno . Sensitivity indices for output on a Riemannian manifold . arXiv e-prints , page arXiv:1810.11591, Oct 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  18. [18]

    Gamboa, P

    F. Gamboa, P. Gremaud, T. Klein, and A. Lagnoux. Supplementary material to ``global sensitivity analysis: A novel generation of mighty estimators based on rank statistics''. Bernoulli , 28(4), 2022

    work page 2022

  19. [19]

    Gamboa, A

    F. Gamboa, A. Janon, T. Klein, A. Lagnoux, and C. Prieur. Statistical inference for S obol P ick- F reeze M onte C arlo method. Statistics , 50(4):881--902, 2016

    work page 2016

  20. [20]

    Gamboa, T

    F. Gamboa, T. Klein, and A. Lagnoux. Sensitivity analysis based on Cram \'e r von Mises distance . SIAM/ASA Journal on Uncertainty Quantification , 6(2):522--548, Apr. 2018

    work page 2018

  21. [21]

    Gamboa, T

    F. Gamboa, T. Klein, and A. Lagnoux. A central limit theorem for generalized L -statistics. Preprint, 2021

    work page 2021

  22. [22]

    Gamboa, T

    F. Gamboa, T. Klein, A. Lagnoux, and L. Moreno. Sensitivity analysis in general metric spaces. Reliability Engineering & System Safety , 212:107611, 2021

    work page 2021

  23. [23]

    J. Hart, P. Gremaud, and T. David. Global sensitivity analysis of high dimensional neuroscience models: an example of neurovascular coupling. Bull Math Biol , 2019

    work page 2019

  24. [24]

    Hoeffding

    W. Hoeffding. A class of statistics with asymptotically normal distribution. Ann. Math. Statistics , 19:293--325, 1948

    work page 1948

  25. [25]

    Janon, T

    A. Janon, T. Klein, A. Lagnoux, M. Nodet, and C. Prieur. Asymptotic normality and efficiency of two S obol index estimators. ESAIM: Probability and Statistics , 18:342--364, 1 2014

    work page 2014

  26. [26]

    Kucherenko and S

    S. Kucherenko and S. Song. Different numerical estimators for main effect global sensitivity indices. Reliability Engineering & System Safety , 165:222--238, 2017

    work page 2017

  27. [27]

    McDiarmid

    C. McDiarmid. On the method of bounded differences. Surveys in combinatorics , 141(1):148--188, 1989

    work page 1989

  28. [28]

    Orey et al

    S. Orey et al. A central limit theorem for m -dependent random variables. Duke Mathematical Journal , 25(4):543--546, 1958

    work page 1958

  29. [29]

    A. B. Owen. Better estimation of small sobol' sensitivity indices. ACM Trans. Model. Comput. Simul. , 23(2):11:1--11:17, may 2013

    work page 2013

  30. [30]

    A. B. Owen. Variance components and generalized S obol' indices. SIAM/ASA Journal on Uncertainty Quantification , 1(1):19--41, 2013

    work page 2013

  31. [31]

    A. B. Owen, J. Dick, and S. Chen. Higher order S obol' indices. Information and Inference , 3(1):59--81, 2014

    work page 2014

  32. [32]

    Saltelli, K

    A. Saltelli, K. Chan, and E. Scott. Sensitivity analysis . Wiley Series in Probability and Statistics. John Wiley & Sons, Ltd., Chichester, 2000

    work page 2000

  33. [33]

    T. J. Santner, B. Williams, and W. Notz. The Design and Analysis of Computer Experiments . Springer-Verlag, 2003

    work page 2003

  34. [34]

    H. Shi, M. Drton, and F. Han. On the power of chatterjee rank correlation, 2020

    work page 2020

  35. [35]

    I. M. Sobol. Sensitivity estimates for nonlinear mathematical models. Math. Modeling Comput. Experiment , 1(4):407--414 (1995), 1993

    work page 1995

  36. [36]

    I. M. Sobol. Global sensitivity indices for nonlinear mathematical models and their M onte C arlo estimates . Mathematics and Computers in Simulation , 55(1-3):271--280, 2001

    work page 2001

  37. [37]

    B. Sudret. Global sensitivity analysis using polynomial chaos expansions . Reliability Engineering & System Safety , 93(7):964--979, 2008

    work page 2008

  38. [38]

    Trutschnig

    W. Trutschnig. On a strong metric on the space of copulas and its induced dependence measure. Journal of mathematical analysis and applications , 384(2):690--705, 2011

    work page 2011

  39. [39]

    A. W. van der Vaart. Asymptotic statistics , volume 3 of Cambridge Series in Statistical and Probabilistic Mathematics . Cambridge University Press, Cambridge, 1998

    work page 1998