pith. sign in

arxiv: 2605.15814 · v1 · pith:BSAWYPEJnew · submitted 2026-05-15 · 🧮 math.ST · stat.TH

Goodness-of-Fit Testing for Point Processes in Large Populations

Sami Umut Can , Estate V. Khmaladze , Roger J. A. Laeven This is my paper

Pith reviewed 2026-05-19 19:42 UTC · model grok-4.3

classification 🧮 math.ST stat.TH
keywords goodness-of-fit testingpoint processesparametric intensityunitary transformationdistribution-free testssurvival processessoftware reliability
0
0 comments X

The pith

A unitary transformation maps the natural testing process for parametric point processes to a limiting target whose distribution is free of unknown intensity parameters.

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

The paper develops a goodness-of-fit procedure for testing whether the conditional intensity of a point process observed in a large population belongs to a specified parametric family. Starting from a natural testing process that depends on the unknown parameters, the authors construct a unitary transformation of that process. Under the null hypothesis the transformed process converges weakly to a fixed standard target process whose law does not involve the parameters at all. This independence makes the resulting tests asymptotically distribution-free, so critical values can be tabulated once and for all. The method is checked in Monte Carlo experiments on Aalen survival processes, mixture cure models and software reliability models, and is applied to real human lifetime data and software failure records.

Core claim

Constructing a unitary transformation of the natural parametric testing process produces a limiting process that converges weakly to a standard target independent of the particular parametric form assumed under the null hypothesis, thereby enabling asymptotically distribution-free goodness-of-fit testing for parametric point processes.

What carries the argument

Unitary transformation of the natural parametric testing process that removes dependence on the unknown parameters in the intensity family.

If this is right

  • Goodness-of-fit tests for point-process intensities can be performed without deriving parameter-specific limiting distributions.
  • The same transformed process applies directly to Aalen-type survival processes with or without censoring.
  • The approach covers mixture cure models and software reliability models with comparable asymptotic behavior.
  • Monte Carlo evidence indicates that the tests maintain reasonable size and power in moderate sample sizes.
  • The procedure can be applied immediately to observed lifetime records and failure-count data.

Where Pith is reading between the lines

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

  • The same transformation idea could be examined for other classes of counting processes whose intensity estimators produce non-pivotal limits.
  • Explicit construction of the unitary operator for common intensity families would allow ready implementation in statistical packages.
  • Power comparisons against local alternatives that perturb the intensity function could be derived from the same transformed process.
  • Links to the theory of empirical processes for marked point processes might yield sharper rates of convergence.

Load-bearing premise

A unitary transformation exists that maps the natural parametric testing process onto a limiting target whose distribution does not depend on the unknown parameters of the intensity family.

What would settle it

For any parametric family, compute the transformed process on simulated data under the null and check whether its finite-dimensional distributions or empirical measure converge to the claimed parameter-free target; persistent dependence on parameters or failure to match the standard limit would refute the claim.

Figures

Figures reproduced from arXiv: 2605.15814 by Estate V. Khmaladze, Roger J. A. Laeven, Sami Umut Can.

Figure 5.1
Figure 5.1. Figure 5.1: Testing for the Aalen model (26) under the null hypothesis. The first row shows the empirical distribution functions of the three test statistics in (29), as computed from Wcn and from Wcµn . The second row shows the empirical distribution functions of the same test statistics, as computed from TcWcn and Wcµn . Unlike the empirical process Wcn, the transformed process TcWcn seems to behave identically to… view at source ↗
Figure 5.2
Figure 5.2. Figure 5.2: Testing for the Jelinski-Moranda model (31) under the null hypothesis. The first row shows the empirical distribution functions of the three test statistics in (29), as computed from Wcn and from Wcµn . The second row shows the empirical distribution functions of the same test statistics, as computed from TcWcn and Wcµn . Unlike the empirical process Wcn, the transformed process TcWcn seems to behave ide… view at source ↗
Figure 5.3
Figure 5.3. Figure 5.3: Testing for the Littlewood model (32) under the null hypothesis. The first row shows the empirical distribution functions of the three test statistics in (29), as computed from Wcn and from Wcµn . The second row shows the empirical distribution functions of the same test statistics, as computed from TcWcn and Wcµn . Unlike the empirical process Wcn, the transformed process TcWcn seems to behave identical… view at source ↗
Figure 5.4
Figure 5.4. Figure 5.4: Testing for the Aalen model (26) under the alternative hypothesis. The first row shows the empirical distribution functions of the three test statistics in (29), as computed from Wcn and from Wcµn . The second row shows the empirical distribution functions of the same test statistics, as computed from TcWcn and Wcµn . In contrast to the situation under the null hypothesis, the transformed process TcWcn i… view at source ↗
Figure 6.1
Figure 6.1. Figure 6.1: The observed process Nn(t)/n (solid line) compared with its estimated limits under the Weibull model (dotted line) and the Gompertz model (dashed line). The plot is split into two periods to make the differences more visible [PITH_FULL_IMAGE:figures/full_fig_p022_6_1.png] view at source ↗
Figure 6.2
Figure 6.2. Figure 6.2: The observed process Nn(t)/n (solid line) compared with its estimated limits under the Jelinski-Moranda model (dotted line) and the Littlewood model (dashed line) Anderson-Darling. It appears that the Littlewood model is a more plausible model for this data set than Jelinski-Moranda. In [PITH_FULL_IMAGE:figures/full_fig_p024_6_2.png] view at source ↗
read the original abstract

Suppose we have an observed path from a point process counting event occurrences in a large population. Based on the observed path, we would like to test the null hypothesis that the conditional intensity of the point process belongs to a particular parametric family. We propose a novel approach to conducting such goodness-of-fit tests. The idea is to construct a unitary transformation of a natural parametric testing process such that it converges weakly to a ``standard'' target process, independent of the particular parametric form assumed under the null hypothesis. This transformation therefore paves the way for asymptotically distribution-free goodness-of-fit testing of parametric point processes. We demonstrate the good finite-sample performance of our approach through Monte Carlo simulations of Aalen-type survival processes, without and with censoring, mixture cure models, and software reliability models, and we illustrate its applicability with observed human lifetimes as well as real software failures.

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

1 major / 2 minor

Summary. The paper proposes a unitary transformation of a natural parametric testing process for observed paths of point processes in large populations. Under the null that the conditional intensity belongs to a given parametric family, the transformed process is claimed to converge weakly to a standard target process whose distribution is independent of the particular parametric form. This is positioned as enabling asymptotically distribution-free goodness-of-fit tests. The approach is supported by Monte Carlo simulations for Aalen-type survival processes (with and without censoring), mixture cure models, and software reliability models, plus illustrations on real human lifetime and software failure data.

Significance. If the central transformation result holds under suitable regularity conditions, the method would supply a useful, distribution-free tool for goodness-of-fit assessment in parametric point-process models arising in survival analysis and reliability. The breadth of simulation settings and the real-data examples provide concrete evidence of practical applicability and finite-sample behavior.

major comments (1)
  1. [theoretical development and main theorem] The abstract asserts that the unitary transformation yields weak convergence to a parameter-free limiting target process. However, the manuscript supplies neither the explicit construction of this transformation (presumably via score or compensator operators) nor a derivation of the weak-convergence result, nor the precise regularity conditions under which the limiting covariance kernel becomes independent of θ. This omission is load-bearing for the central claim, especially for the Aalen-type models with censoring and the mixture cure models included in the simulations, where the compensator satisfies an integral equation whose orthogonal complement may retain θ-dependence absent an additional orthogonality condition.
minor comments (2)
  1. [Section 2] The description of the natural parametric testing process (prior to transformation) should include its explicit definition and the precise form of the intensity estimator used under the null.
  2. [simulation studies] Finite-sample diagnostics (e.g., empirical coverage of critical values or QQ plots of the transformed statistic) would strengthen the simulation section; currently only qualitative statements of “good performance” are given.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address the major comment point by point below and will incorporate revisions to strengthen the theoretical presentation.

read point-by-point responses

  1. Referee: [theoretical development and main theorem] The abstract asserts that the unitary transformation yields weak convergence to a parameter-free limiting target process. However, the manuscript supplies neither the explicit construction of this transformation (presumably via score or compensator operators) nor a derivation of the weak-convergence result, nor the precise regularity conditions under which the limiting covariance kernel becomes independent of θ. This omission is load-bearing for the central claim, especially for the Aalen-type models with censoring and the mixture cure models included in the simulations, where the compensator satisfies an integral equation whose orthogonal complement may retain θ-dependence absent an additional orthogonality condition.

    Authors: We agree that the explicit construction of the unitary transformation and the full derivation of the weak-convergence result require more detailed exposition to support the central claim. In the revised manuscript we will insert a dedicated subsection that gives the explicit form of the transformation in terms of the score operator and the compensator, followed by a complete proof of weak convergence to the parameter-free target process. We will also state the precise regularity conditions (including the required orthogonality with respect to the parametric family) that guarantee the limiting covariance kernel is independent of θ. For the Aalen-type models with censoring and the mixture cure models, we will verify that the integral equation satisfied by the compensator admits the necessary orthogonality condition and will add this verification explicitly. These changes will be placed in the main theoretical section with supporting technical details moved to an appendix if needed. revision: yes

Circularity Check

0 steps flagged

No circularity: unitary transformation is a constructive proposal, not a reduction to inputs

full rationale

The paper's central claim is the existence and application of a unitary transformation that maps a natural parametric testing process to a weakly convergent limit independent of the intensity parameter. This is presented as a novel construction enabling distribution-free tests, with demonstrations via Monte Carlo simulations on Aalen models, censoring, mixture cure models, and software reliability. No quoted equations or steps reduce the claimed limit or transformation to a fitted quantity, self-definition, or self-citation chain; the derivation remains self-contained as an explicit operator construction whose independence property is asserted and verified externally through simulation rather than by tautological re-expression of the null family.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger records the high-level modeling assumptions stated there. The approach relies on standard weak-convergence results for point processes but introduces a new transformation whose technical justification is not shown.

axioms (1)
  • domain assumption The observed path comes from a point process whose conditional intensity belongs to a specified parametric family under the null hypothesis.
    This is the explicit setup for the goodness-of-fit test described in the first sentence of the abstract.

pith-pipeline@v0.9.0 · 5680 in / 1367 out tokens · 59942 ms · 2026-05-19T19:42:32.873412+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

70 extracted references · 70 canonical work pages

  1. [1]

    2002 , PUBLISHER =

    The Statistical Analysis of Failure Time Data , AUTHOR =. 2002 , PUBLISHER =

    work page 2002

  2. [2]

    1996 , PUBLISHER =

    Weak Convergence and Empirical Processes: With Applications to Statistics , AUTHOR =. 1996 , PUBLISHER =

    work page 1996

  3. [3]

    2007 , PUBLISHER =

    Random Fields and Geometry , AUTHOR =. 2007 , PUBLISHER =

    work page 2007

  4. [4]

    and Wei, L.-J

    Lin, D. and Wei, L.-J. and Ying, Z. , JOURNAL =. Checking the

    work page

  5. [5]

    and Dobler, Dennis and de Gunst, Mathisca C

    Dietrich, Marina T. and Dobler, Dennis and de Gunst, Mathisca C. M. , JOURNAL =. Wild bootstrap for counting process‑based statistics:

    work page

  6. [6]

    Dynamic Regression Models for Survival Data , AUTHOR =

    work page

  7. [7]

    Statistical Inference for a Family of Counting Processes , AUTHOR =

    work page

  8. [8]

    and Can, S

    Baars, J. and Can, S. U. and Laeven, R. J. A. , title =

    work page

  9. [9]

    Stochastic Process

    Limit theorems for Markovian Hawkes processes with a large initial intensity , AUTHOR =. Stochastic Process. Appl. , VOLUME =

    work page

  10. [10]

    Integrability of expected increments of point processes and a related random change of scale , AUTHOR =. Trans. Amer. Math. Soc. , VOLUME =

    work page

  11. [11]

    A simple proof of the multivariate random time change theorem for point processes , AUTHOR =. J. Appl. Probab. , VOLUME =

    work page

  12. [12]

    Random time changes for multivariate counting processes , AUTHOR =. Scand. Actuar. J. , VOLUME =

    work page

  13. [13]

    A Martingale Approach to Point Processes , AUTHOR =

    work page

  14. [14]

    Meyer, Paul-Andr. D. S

    work page

  15. [15]

    Statistical Models Based on Counting Processes , AUTHOR =

    work page

  16. [16]

    Statistical Methods with Applications to Demography and Life Insurance , AUTHOR =

    work page

  17. [17]

    Nonparametric inference for a family of counting processes , AUTHOR =. Ann. Statist. , VOLUME =

    work page

  18. [18]

    Point Processes and Their Statistical Inference , AUTHOR =

    work page

  19. [19]

    An Introduction to the Theory of Point Processes: Volumes I and II , AUTHOR =

    work page

  20. [20]

    Bernoulli , VOLUME =

    Unitary transformations, empirical processes and distribution free testing , AUTHOR =. Bernoulli , VOLUME =

    work page

  21. [21]

    Aalen, Odd O. and. Survival and Event History Analysis:

    work page

  22. [22]

    and Harrington, David P

    Fleming, Thomas R. and Harrington, David P. , publisher =

    work page

  23. [23]

    Contemporary Multivariate Analysis and Design of Experiments , PAGES =

    Multidimensional variation for quasi-Monte Carlo , AUTHOR =. Contemporary Multivariate Analysis and Design of Experiments , PAGES =

    work page

  24. [24]

    Sample moduli for set-indexed

    Alexander, Kenneth S , JOURNAL =. Sample moduli for set-indexed

    work page

  25. [25]

    2004 , PUBLISHER =

    Statistics of Extremes: Theory and Applications , AUTHOR =. 2004 , PUBLISHER =

    work page 2004

  26. [26]

    , YEAR =

    Billingsley, P. , YEAR =. Convergence of

    work page

  27. [27]

    2001 , PUBLISHER =

    An Introduction to Statistical Modeling of Extreme Values , AUTHOR =. 2001 , PUBLISHER =

    work page 2001

  28. [28]

    2000 , PUBLISHER =

    Extreme Value Distributions: Theory and Applications , AUTHOR =. 2000 , PUBLISHER =

    work page 2000

  29. [29]

    and Tawn, Jonathan A

    Coles, Stuart G. and Tawn, Jonathan A. , TITLE =. J. Roy. Statist. Soc. Ser. B , VOLUME =. 1991 , NUMBER =

    work page 1991

  30. [30]

    Limit theory for multivariate sample extremes , AUTHOR =. Z. Wahrsch. verw. Gebiete , VOLUME =

    work page

  31. [31]

    Delgado, M. A. and Hidalgo, J. and Velasco, C. , TITLE =. Ann. Statist. , VOLUME =. 2005 , PAGES =

    work page 2005

  32. [32]

    and Hetzler, B

    Dette, H. and Hetzler, B. , TITLE =. Math. Methods Statist. , VOLUME =. 2009 , PAGES =

    work page 2009

  33. [33]

    Approximations to the tail empirical distribution function with application to testing extreme value conditions , AUTHOR =. J. Statist. Plann. Inference , VOLUME =

    work page

  34. [34]

    and de Haan, L

    Einmahl, J.H.J. and de Haan, L. and Sinha, A.K. , TITLE =. Stochastic Process. Appl. , VOLUME =. 1997 , PAGES =

    work page 1997

  35. [35]

    Bernoulli , VOLUME =

    A method of moments estimator of tail dependence , AUTHOR =. Bernoulli , VOLUME =

    work page

  36. [36]

    An M-estimator for tail dependence in arbitrary dimensions , AUTHOR =. Ann. Statist. , VOLUME =

    work page

  37. [37]

    Weighted approximations of tail copula processes with application to testing the bivariate extreme value condition , AUTHOR =. Ann. Statist. , VOLUME =

    work page

  38. [38]

    Parametric tail copula estimation and model testing , AUTHOR =. J. Multivariate Anal. , VOLUME =

    work page

  39. [39]

    and Weissman, Ishay , TITLE =

    Joe, Harry and Smith, Richard L. and Weissman, Ishay , TITLE =. J. Roy. Statist. Soc. Ser. B , VOLUME =. 1992 , NUMBER =

    work page 1992

  40. [40]

    , TITLE =

    Khmaladze, Estate V. , TITLE =. Theory Probab. Appl. , VOLUME =. 1981 , PAGES =

    work page 1981

  41. [41]

    , TITLE =

    Khmaladze, Estate V. , TITLE =. Ann. Statist. , VOLUME =. 1988 , PAGES =

    work page 1988

  42. [42]

    , TITLE =

    Khmaladze, Estate V. , TITLE =. Ann. Statist. , VOLUME =. 1993 , PAGES =

    work page 1993

  43. [43]

    Econometrica , VOLUME =

    Inference on the quantile regression process , AUTHOR =. Econometrica , VOLUME =

    work page

  44. [44]

    Quantile autoregression , AUTHOR =. J. Amer. Statist. Assoc. , VOLUME =

    work page

  45. [45]

    , JOURNAL =

    Krause, J.M. , JOURNAL =. Fouriersche

    work page

  46. [46]

    Model checks under random censorship , AUTHOR =. Statist. Probab. Lett. , VOLUME =

    work page

  47. [47]

    and Thies, S

    Stute, W. and Thies, S. and Zhu, L. , TITLE =. Ann. Statist. , VOLUME =. 1998 , PAGES =

    work page 1998

  48. [48]

    , TITLE =

    Towghi, N. , TITLE =. J. Inequal. Pure Appl. Math. , VOLUME =

    work page

  49. [49]

    Martingale transforms goodness-of-fit tests in regression models , AUTHOR =. Ann. Statist. , VOLUME =

    work page

  50. [50]

    Goodness-of-fit problem for errors in nonparametric regression: Distribution free approach , AUTHOR =. Ann. Statist. , VOLUME =

    work page

  51. [51]

    and Nikabadze, A

    McKeague, Ian W. and Nikabadze, A. M. and Sun, Yan Qing , TITLE =. Ann. Statist. , VOLUME =. 1995 , NUMBER =

    work page 1995

  52. [52]

    Estimating the limit distribution of multivariate extremes , AUTHOR =. Comm. Statist. Stochastic Models , VOLUME =

    work page

  53. [53]

    Hardy, G. H. , JOURNAL =. On double

    work page

  54. [54]

    Topological algebras of functions of bounded variation

    Bl\". Topological algebras of functions of bounded variation. Manuscripta Math. , VOLUME =

    work page

  55. [55]

    , YEAR =

    Huang, X. , YEAR =. Statistics of

    work page

  56. [56]

    and Ferreira, A

    de Haan, L. and Ferreira, A. , YEAR =. Extreme

    work page

  57. [57]

    Introduction to the

    Hildebrandt, Theophil Henry , YEAR =. Introduction to the

    work page

  58. [58]

    Multivariate Analysis V , YEAR =

    Bivariate extremes: Foundations and statistics , AUTHOR =. Multivariate Analysis V , YEAR =

    work page

  59. [59]

    International Encyclopedia of Statistical Science , YEAR =

    Khmaladze transformation , AUTHOR =. International Encyclopedia of Statistical Science , YEAR =

    work page

  60. [60]

    Some notes on goodness-of-fit tests and innovation martingales , AUTHOR =. Proc. A. Razmadze Math. Inst. , VOLUME =

    work page

  61. [61]

    Supplement to ``Asymptotically Distribution-Free Goodness-of-Fit Testing for Tail Copulas'' , AUTHOR =

    work page

  62. [62]

    Statistical Computer Performance Evaluation , publisher =

    SOFTWARE RELIABILITY RESEARCH , editor =. Statistical Computer Performance Evaluation , publisher =. 1972 , author =

    work page 1972

  63. [63]

    Simulations on the

    van. Simulations on the. Stat. Comput. , VOLUME =. 1992 , PAGES =

    work page 1992

  64. [64]

    , journal=

    Littlewood, B. , journal=. Theories of software reliability:. 1980 , volume=

    work page 1980

  65. [65]

    Ross Maller and Sidney Resnick and Soudabeh Shemehsavar and Muzhi Zhao , title =. Stat. Surv. , pages =

    work page

  66. [66]

    , title =

    Boag, John W. , title =. J. R. Stat. Soc. Ser. B Stat. Method. , volume =

    work page

  67. [67]

    1996 , isbn =

    Handbook of. 1996 , isbn =

    work page 1996

  68. [68]

    2025 , howpublished =

    Human Mortality Database , author =. 2025 , howpublished =

    work page 2025

  69. [69]

    Tiwari and Jyoti N

    Yanqing Sun and Ram C. Tiwari and Jyoti N. Zalkikar , journal =. Goodness of Fit Tests for Multivariate Counting Process Models with Applications , volume =

    work page

  70. [70]

    Jelinski-

    Lutfiah Al Turk and Eftekhar Gabel Alsolami , journal=. Jelinski-. 2016 , volume=

    work page 2016