Pith. sign in

REVIEW 3 major objections 4 minor 36 references

Post Hoc Inference for Component Attribution in Multivariate Change-Point Detection

T0 review · 3 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read After a change-point is found, two tests identify which blocks of variables caused it while controlling false attribution.

desk verdict A useful framework for post hoc block attribution with valid guarantees under explicit localization assumptions, but the implemented version swaps in an empirically calibrated constant that is not proven to satisfy the assumptions, so the finite-sample level claim is conditional on unverified calibration. read the letter →

arxiv 2607.14814 v1 pith:XUQTC2O6 submitted 2026-07-16 stat.ME math.STstat.MLstat.TH

Dhia-Elhaq Ouerfelli , Sylvain Arlot , Kevin Bleakley , Patrick Pamphile This is my paper
classification stat.MEmath.STstat.MLstat.TH MSC 62G1062H1562M10
keywords posthocchange-pointattributionmultivariatetimeseriestwo-sampletestingkernelMMDTypeIerrorcontrolhold-outvalidationgrid-basedselectionbias
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

After an offline change-point detector finds a break in a multivariate time series, the analyst usually wants to know which coordinates drove it. This paper proves that two procedures — a grid-based test on the full series (GTST) and a hold-out split — control the probability of falsely blaming a block. Both procedures use a generic two-sample test (for instance, a kernel MMD test) on segments that are deliberately kept away from the uncertain estimated boundary. The stated guarantee is a bound α0 + α1: α0 for the chance the initial detector mislocalizes, α1 for the local test's false-alarm rate. The paper also shows, in simulations, that a naive test applied directly at the estimated change-point inflates false positives severely.

What carries the argument

GTST enumerates candidate pairs of pre/post segments on a grid of step η inside the δn-neighborhood of the estimated break, applies the chosen two-sample test to every pair, and rejects only if every local test rejects; this intersection device lets the paper avoid conditioning on a single random segment. The hold-out method assigns detection to a subset Idet and inference to the complement, so the test statistic is independent of the estimated change-point. The uncertainty radius δn and the minimal-spacing condition nΛτ⋆ > δ′n are what guarantee that the selected segments contain i.i.d. observations from a single distribution.

What would settle it

Simulate weak mean shifts with small n, estimate δn exactly as in the paper, and record the empirical frequency of |τ̂ − τ*| > δn; if that frequency exceeds α0, then Assumption 1 is violated and the GTST Type I error can be checked to see whether it exceeds α0 + α1.

Watch

Extended reading notes

Core claim

The central claim is Theorem 3 and Theorem 9: under a localization event of probability at least 1 − α0 and a minimal spacing condition, GTST and the hold-out test are valid level-α0 + α1 tests of whether a predefined block's distribution changes at the candidate change-point. Validity comes from discarding an uncertainty window of width δn around each estimated change-point; on the high-probability localization event, the remaining segments lie inside true stationary segments, so the i.i.d. assumption of the local two-sample test is restored. GTST additionally checks a separability condition and, if it fails, conservatively retains the null. The hold-out instead runs detection on one subset

Load-bearing premise

A detection algorithm must output the correct number of change-points and localize them within δn with probability at least 1 − α0, and in the implemented version δn is computed with a calibrated constant rather than the proven bound — if that radius is too small, the guarantee collapses.

Editorial extensions

If this is right

  • Any two-sample test with a valid level can be plugged into GTST or the hold-out without re-deriving the main guarantee.
  • For two blocks, the family-wise error rate is controlled at α0 + 2α1; for L blocks, at α0 + Lα1 via a union bound.
  • In the single-change-point case, the hold-out test attains level α1 rather than α0 + α1, because inference data are independent of detection.
  • The framework is nonparametric and applies to variance and covariance changes, not just mean shifts, as shown in simulations.
  • The exact block recovery rate approaches oracle power once the signal is strong enough, while naive testing produces frequent false attributions.

Reading between the lines

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

  • If the guarantees are correct, the common practice of testing directly on data-driven segments is not merely risky but theoretically invalid; the simulations quantify the inflation, with naive Type I error reaching about 0.25 for a small mean shift.
  • The hold-out argument suggests a natural extension to online change-point monitoring: detect on a buffer, then test on fresh observations, preserving valid attribution without needing a localization bound.
  • Because the empirical calibration of the localization radius uses a tuned constant, users should report the observed distribution of localization errors, not just the calibrated δn, to verify Assumption 1 in practice.
  • For weak signals where GTST's separability condition often fails, the hold-out is the safer default; once the signal is moderate, GTST is more powerful.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes two post hoc procedures, GTST and hold-out, for attributing a detected multivariate change-point to one of two predefined coordinate blocks. The GTST procedure builds a grid of candidate segment pairs around estimated change-points and rejects the block-level null only if every local two-sample test rejects; the hold-out procedure splits the series into detection and inference sets and performs the local test only on the inference set. The main theoretical results, Theorems 3 and 9, claim that each block-level test controls the Type I error at level α0 + α1, where α0 is the localization failure probability and α1 is the nominal level of the local two-sample test, under Assumptions 1–5. The empirical section instantiates the local tests as MMD permutation tests and computes the operational localization radius δn using a calibrated constant C0 = 0.15 in place of the proven constant 148 from the Garreau–Arlot bound.

Significance. If the operational gap identified below is closed, the paper makes a useful contribution to post-selection inference for change-point attribution. The conditional results are clean: the error decomposition into a localization event and a two-sample testing event is elegant, and the hold-out treatment in the single-change-point case is correctly shown to require only independence, not localization. The grid-based product rule is a conservative but theoretically transparent device. The manuscript also provides code and reproducible simulation code, which is a strength. The central weakness is that the implemented algorithm does not provably satisfy the assumptions under which the theorems are stated.

major comments (3)
  1. [Section 4.3.1, Algorithm 1 (definition of δn)] The operational procedure replaces the proven constant 148 in the Garreau–Arlot localization bound with an empirically calibrated C0 = 0.15 and computes δn = C0 κ⋆ M² / Δ̂² (log(1/α0) + log n + 1), using a plug-in estimate Δ̂ of the minimal MMD jump. No theorem shows that this data-dependent δn satisfies P(d∞(τ⋆, τ̂) > δn) ≤ α0. Theorems 3 and 9 are explicitly conditional on Assumption 1, which requires a δn satisfying exactly this high-probability localization event. Because δn is random and may under-estimate the true localization error, the event Ωα0 may fail with probability greater than α0, and the level α0 + α1 is not guaranteed for the procedure as actually run in the simulations and real-data experiment. Appendix C is a finite calibration study over selected distribution families and dimensions; it does not establish a worst-case bound. This is load-bearing because the paper's ce
  2. [Algorithm 1, Step 3; Remark 5; Remark 7] The separability condition is checked using the plug-in estimate Λ̂τ, but Assumption 2 requires the true minimal spacing nΛτ⋆ > δ′n. If Λ̂τ over-estimates the true spacing, the procedure executes the multi-test grid despite being outside the regime of Lemma 1 and Theorem 3. The paper correctly notes in Remark 5 that Condition (2.3) cannot be checked directly, but it does not provide a valid alternative. For the hold-out method, Remark 7 explicitly acknowledges that if the spacing condition is violated, 'formal guarantees are lost' while the procedure is still applied. This means the theoretical level control applies only to an idealized version of the algorithms in which the true parameters are known; the practical algorithms need either a valid data-dependent check or a clearly stated limitation that the guarantee is conditional on a condition that is not verified.
  3. [Abstract and Section 5] The abstract and conclusion state that the proposed procedures 'provide theoretical guarantees for Type I error control' without qualifying that the guarantees hold only under Assumptions 1–5, which are not established for the implemented operational pipeline using C0 = 0.15 and plug-in estimates. The statements should be softened to describe the guarantees as conditional on a valid localization event, or the operational procedure should be modified to use a proven conservative δn so that the unconditional claim is justified for the implemented algorithm.
minor comments (4)
  1. [Section 3.1.2, Lemma 2] In Lemma 2, point 3, the restricted grid-set is denoted eE_i0,η (with a tilde) in the statement but the definition immediately before uses eEi0,η without the tilde. The notation should be harmonized.
  2. [Section 4.3.1] The displayed formula for the theoretical condition has 'log(1/α0)+log n+1' in the numerator of a fraction that is not explicitly parenthesized. It should be written as Δ²/M² > 148 κ⋆(κ⋆+1)(log(1/α0)+log n+1)/(nΛτ⋆) to avoid ambiguity.
  3. [Remark 10] The statement that the permutation test provides an exact non-asymptotic level α1 is correct only when ties are handled appropriately; the reference [27] is about stepdown methods and not specifically about permutation MMD tests. A more direct reference, such as the standard permutation testing literature, would be preferable.
  4. [Assumption 4] Assumption 4 uses the notation bκdet before it is defined; it should be introduced as κ̂det for clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the level guarantees are conditional derivations from stated localization/spacing assumptions; the empirical C0 calibration is an acknowledged validity gap, not a fitted prediction of the attribution outcome.

full rationale

I walked the derivation chain. Theorem 3 and Theorem 9 are direct probabilistic decompositions: P(reject Hφ0) ≤ P(Ωα0^c) + P(the generic level-α1 test rejects on the oracle projected pair) ≤ α0 + α1, using Lemma 2 (GTST) and Lemma 8 (hold-out) to guarantee that under Hφ0 the relevant observations are i.i.d. No fitted parameter enters these derivations: the grid step, the product rule, and the levels α0, α1 are fixed before the analysis, and the localization input is the existence of an event Ωα0 supplied by Assumption 1 (or Assumption 4 for the detection set). The proof does not define the test in terms of its own target result; the projected oracle segments are shown, by the paper's own equations, to belong to the candidate collection. The reliance on Garreau & Arlot (2018), including co-author Arlot, is a self-citation, but it is a published theorem with explicit assumptions and an explicit constant (148), not an unverified uniqueness claim or an ansatz smuggled in via citation. The genuine weakness is operational: in Section 4.3.1 the paper replaces 148 by an empirically calibrated C0 = 0.15 and plugs in estimates of κ⋆, Δ and Λτ⋆, so the computed δn is not proven to satisfy Assumption 1. The paper itself acknowledges this in Appendix C: 'Even when replacing 148 by the empirical value of C0 obtained from the simulations, the direct application of Theorem 11 remains difficult. Indeed, several parameters involved in the definition of C0 are typically unknown in practice.' That is a correctness/robustness gap, not a circular reduction: the simulated Type I error is not mathematically forced by the calibration, and the central theorems remain self-contained conditional statements. Under the review rules, this non-finding is appropriate: no 'prediction' is equivalent to its fitted inputs by construction.

Assumptions & free parameters 3 free parameters · 8 assumptions · 0 invented entities

The main load-bearing inputs are the localization and spacing assumptions (Assumptions 1–5), plus the empirical calibration C0. No new objects (particles, forces, etc.) are introduced. The contributions are procedural.

free parameters (3)
  • C0 = 0.15 (default; 0.08–0.15 depending on dimension/κ*)
    Replaces the proven constant 148 in the localization bound of Garreau & Arlot (Thm 11). Calibrated empirically over simulation configurations in Appendix C; the operational δn and hence the practical Type I guarantee depend on it.
  • grid step η = max(⌊δn/2⌋, 1) in experiments
    User-selected discretization parameter controlling the number of candidate segments; validity holds for any η, but power and cost vary.
  • kernel bandwidth σ (median heuristic) = data-dependent median pairwise distance
    Standard heuristic used for the Gaussian kernel in both KCP and MMD tests; not a target-result fit.
assumptions (8)
  • domain assumption Assumption 1: correct number of change-points and localization error ≤δn on an event of probability ≥1−α0
    Central premise for both GTST (Thm 3) and hold-out (Thm 9). The whole framework depends on the upstream detector satisfying this; in practice δn is computed with the calibrated C0.
  • domain assumption Assumption 2: minimal spacing nΛτ* > 2δn (κ*>2) or > δn (κ*=2)
    Ensures separated confidence intervals and nonempty oracle-projected segments, giving GTST power; without it the test is conservative.
  • domain assumption Assumption 3: minimal coverage condition for hold-out (every true segment contains at least one detection index)
    Needed so the detection set induces a valid segmentation; may fail for short segments with odd/even splits.
  • domain assumption Assumption 4: localization on the detection set
    Mirror of Assumption 1 for the hold-out method.
  • domain assumption Assumption 5: minimal spacing on the detection set
    Mirror of Assumption 2 for the hold-out method.
  • domain assumption Observations are mutually independent and i.i.d. within each true segment
    Standard change-point model; required for the two-sample tests to be valid.
  • domain assumption Kernel is bounded and characteristic
    Needed for the MMD threshold test and for guaranteed metric property.
  • ad hoc to paper Empirical constant C0=0.15 majorizes the localization constant for the tested distribution families in dimension ≥2
    Not a proven theorem; it is an empirical claim from Appendix C. The operational Type I control of the implemented procedure relies on it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Post Hoc Inference for Component Attribution in Multivariate Change-Point Detection." pith.science (2026). https://pith.science/paper/XUQTC2O6

@misc{pith2026260714814,
  author       = {Pith},
  title        = {Pith review of: Post Hoc Inference for Component Attribution in Multivariate Change-Point Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XUQTC2O6}},
  note         = {Machine review of arXiv:2607.14814}
}
read the original abstract

We consider the post-detection analysis of change-points for multivariate time series, with the goal of identifying which coordinates are responsible for a detected change. After a change-point has been located by an offline detection algorithm, we propose post hoc statistical procedures to determine whether the change occurs in either of two predefined blocks of coordinates or in both. Our methods rely on two-sample testing procedures with a particular focus on nonparametric tests; we provide theoretical guarantees for Type I error control. Simulations and a real-data experiment demonstrate the strong performance of the proposed procedures.

Figures

Figures reproduced from arXiv: 2607.14814 by the authors.

Figure 1
Figure 1. Complete workflow of the GTST procedure. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the minimal coverage condition: in a valid partition (top), the condition is met: [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Partition of the detection set Idet: due to the minimal coverage condition, each true segment (e.g., S ⋆ r ) contains at least one detection index (blue dots). The projections π +(τ ⋆ r + 1) and π −(τ ⋆ r+1) respectively identify the first and last available detection indices inside the true segment S ⋆ r . These projections effectively partition Idet into disjoint sub-collections I (r ) det exactly aligned with the… view at source ↗
Figures from the paper (16 more)
Figure 4
Figure 4. Figure 4: Complete illustration of the odd/even hold-out splitting strategy defining the test segments. [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Simulation pipeline and post hoc methods [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]
Figure 6
Figure 6. Figure 6: Performance of the post hoc procedures as a function of the mean jump magnitude. (a) Empirical power with the probability that Condition (2.3) holds (dashed line); (b) empirical Type I error rate; (c) exact block recovery rate. The dashed line in (a) indicates the freq…
Figure 7
Figure 7. Figure 7: Performance of the post hoc procedures as a function of the relative size of the smallest segment Λτ ⋆ , which represents the location of the change-point. The mean jump amplitude is fixed at 0.3. (a) Empirical power, along with the empirical probability of satisfying …
Figure 8
Figure 8. Figure 8: Multivariate time series of household electric power consumption from January 14 (20:00) to [PITH_FULL_IMAGE:figures/full_fig_p025_8.png]
Figure 9
Figure 9. Figure 9: C0 as a function of the dimension of the time series. This defines the curve of C0, evaluated as a function of the minimal discrepancy between distributions, measured through the minimum MMD between consecutive segments. 5. Estimation of C0,sup. For each configuration,…
Figure 10
Figure 10. Figure 10: C0 as a function of Λτ ⋆ [PITH_FULL_IMAGE:figures/full_fig_p035_10.png]
Figure 11
Figure 11. Figure 11: C0 as a function of Λτ ⋆ , with κτ ⋆ = 3. We restrict the computation of these values to time series of lengths n ∈ {100, 500, 1000}, with a single change-point and distributions belonging to P. The fixed parameters are d, κτ ⋆ , and y , while the non-fixed parameters…
Figure 12
Figure 12. Figure 12: C0 as a function of Λτ ⋆ , with κτ ⋆ = 4. C.3 Comments and conclusions In this study, we explored several configurations in order to obtain an empirical value of C0 that is close to its theoretical value. The largest values of C0 obtained in the simulations, around 0.…
Figure 13
Figure 13. Figure 13: Performance of the post hoc procedures as a function of the signal sparsity. The total dimension is fixed at d = 50 and the mean jump amplitude is set to 0.3. The x-axis represents the number of affected coordinates s within the active block. As s increases, the signa…
Figure 14
Figure 14. Figure 14: Performance of the post hoc procedures as a function of the dimension d. The proportion of affected coordinates is maintained at approximately 60% (⌊0.6 d⌋) across all tested dimensions, corresponding to a dense signal regime. (a) Empirical power; (b) Empirical Type I…
Figure 15
Figure 15. Figure 15: Performance of the post hoc procedures as a function of the new variance on the affected coordinates. The pre-change variance is exactly 1.0. (a) Empirical power; (b) Empirical Type I error level. (a) Power (b) Type I error level [PITH_FULL_IMAGE:figures/full_fig_p03…
Figure 16
Figure 16. Figure 16: Performance of the post hoc procedures as a function of the covariance jump magnitude. (a) Empirical power; (b) Empirical Type I error level. For variance changes, the pre-change distribution has unit variance in all dimensions. After the change￾point, the variance of…
Figure 17
Figure 17. Figure 17: Performance of the post hoc procedures as a function of the time series length n. A single change-point is located at the center of the sequence, with a fixed mean jump magnitude of 0.3. (a) Empirical power; (b) Empirical Type I error level. D.2 Estimation / Noise D.2…
Figure 18
Figure 18. Figure 18: Performance of the post hoc procedures as a function of the mean jump magnitude in a three￾segment scenario (κ ⋆ = 3). The sequence contains two true change-points, and inference is specifically performed on the first one (i0 = 1). (a) Empirical power; (b) Empirical T…
Figure 19
Figure 19. Figure 19: Performance of the post hoc procedures as a function of the mean jump magnitude in a four￾segment scenario (κ ⋆ = 4). The sequence contains three true change-points, and inference is explicitly focused on the middle one (i0 = 2). (a) Empirical power; (b) Empirical Typ…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

36 extracted references · 7 linked inside Pith

  1. [1]

    A survey of methods for time series change point detection

    Samaneh Aminikhanghahi and Diane Cook. A survey of methods for time series change point detection. Knowledge and Information Systems, 51, 05 2017

    2017

  2. [2]

    A kernel multiple change-point algorithm via model selection.Journal of Machine Learning Research, 20(162):1–56, 2019

    Sylvain Arlot, Alain Celisse, and Zaid Harchaoui. A kernel multiple change-point algorithm via model selection.Journal of Machine Learning Research, 20(162):1–56, 2019

    2019

  3. [3]

    Prentice Hall, 04 1993

    Michèle Basseville and Igor Nikiforov.Detection of Abrupt Change: Theory and Application, volume 15. Prentice Hall, 04 1993

    1993

  4. [4]

    Springer US, 2004

    Alain Berlinet and Christine Thomas-Agnan.Reproducing Kernel Hilbert Spaces in Probability and Statistics. Springer US, 2004

    2004

  5. [5]

    Theoretical guarantees for change localization using conformal p-values.arXiv preprint arXiv:2510.08749, 2025

    Swapnaneel Bhattacharyya and Aaditya Ramdas. Theoretical guarantees for change localization using conformal p-values.arXiv preprint arXiv:2510.08749, 2025

    arXiv 2025

  6. [6]

    Post-selection inference for quantifying uncertainty in changes in variance.arXiv preprint arXiv:2405.15670, 2024

    Rachel Carrington and Paul Fearnhead. Post-selection inference for quantifying uncertainty in changes in variance.arXiv preprint arXiv:2405.15670, 2024

    arXiv 2024

  7. [7]

    Improving power by conditioning on less in post-selection inference for changepoints.Statistics and Computing, 35(1):8, 2025

    Rachel Carrington and Paul Fearnhead. Improving power by conditioning on less in post-selection inference for changepoints.Statistics and Computing, 35(1):8, 2025. 42

    2025

  8. [8]

    Inferenceinhigh-dimensionalonlinechangepoint detection.Journal of the American Statistical Association, 119(546):1461–1472, 2024

    YudongChen, TengyaoWang, andRichardJ.Samworth. Inferenceinhigh-dimensionalonlinechangepoint detection.Journal of the American Statistical Association, 119(546):1461–1472, 2024. PMID: 38974186

    2024

Show all 36 references
  1. [9]

    D. R. Cox. A note on data-splitting for the evaluation of significance levels.Biometrika, 62(2):441–444, 08 1975

    1975

  2. [10]

    Online centralized non-parametric change-point detection via graph-based likelihood-ratio estimation.arXiv preprint arXiv:2301.03011, 2023

    Alejandro de la Concha, Argyris Kalogeratos, and Nicolas Vayatis. Online centralized non-parametric change-point detection via graph-based likelihood-ratio estimation.arXiv preprint arXiv:2301.03011, 2023

    2023 arXiv

  3. [11]

    Data-driven distributed collaborative fault detection and isolation for large-scale dynamic processes in simultaneous-fault cases.Control Engineering Practice, 141:105718, 2023

    Shuyu Ding, Dingguo Liang, Xiaolu Chen, and Ying Yang. Data-driven distributed collaborative fault detection and isolation for large-scale dynamic processes in simultaneous-fault cases.Control Engineering Practice, 141:105718, 2023

    2023

  4. [12]

    Consistent change-point detection with kernels.Electronic Journal of Statistics, 12(2), 1 2018

    Damien Garreau and Sylvain Arlot. Consistent change-point detection with kernels.Electronic Journal of Statistics, 12(2), 1 2018

    2018

  5. [13]

    Large sample analysis of the median heuristic.arXiv preprint arXiv:1707.07269, 2017

    Damien Garreau, Wittawat Jitkrittum, and Motonobu Kanagawa. Large sample analysis of the median heuristic.arXiv preprint arXiv:1707.07269, 2017

    2017 arXiv

  6. [14]

    Chapman and Hall/CRC, 2021

    Christophe Giraud.Introduction to high-dimensional statistics. Chapman and Hall/CRC, 2021

    2021

  7. [15]

    Borgwardt, Malte J

    Arthur Gretton, Karsten M. Borgwardt, Malte J. Rasch, Bernhard Schölkopf, and Alexander Smola. A kernel two-sample test.Journal of Machine Learning Research, 13(1):723–773, 2012

    2012

  8. [16]

    Retrospective multiple change-point estimation with kernels

    Zaid Harchaoui and Olivier Cappé. Retrospective multiple change-point estimation with kernels. In2007 IEEE/SP 14th Workshop on Statistical Signal Processing, pages 768–772. IEEE, 2007

    2007

  9. [17]

    Individual household electric power consumption, 2012

    Georges Hébrail and Alice Bérard. Individual household electric power consumption, 2012

    2012

  10. [18]

    A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics, 6(2):65–70, 1979

    Sture Holm. A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics, 6(2):65–70, 1979

    1979

  11. [19]

    Testing for a change in mean after changepoint detection.Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(4):1082–1104, 2022

    Sean Jewell, Paul Fearnhead, and Daniela Witten. Testing for a change in mean after changepoint detection.Journal of the Royal Statistical Society Series B: Statistical Methodology, 84(4):1082–1104, 2022

    2022

  12. [20]

    Lehmann and Joseph P

    Erich L. Lehmann and Joseph P. Romano.Testing Statistical Hypotheses. Springer, New York, 3rd edition, 2005

    2005

  13. [21]

    Volatility spillover among sector equity returns under structural breaks.Review of Quantitative Finance and Accounting, 58(3):1063–1080, 04 2022

    Farooq Malik. Volatility spillover among sector equity returns under structural breaks.Review of Quantitative Finance and Accounting, 58(3):1063–1080, 04 2022

    2022

  14. [22]

    Matteson and Nicholas A

    David S. Matteson and Nicholas A. James. A nonparametric approach for multiple change point analysis of multivariate data.Journal of the American Statistical Association, 109(505):334–345, 2014

    2014

  15. [23]

    McMillan and Mark E

    David G. McMillan and Mark E. Wohar. Structural breaks in volatility: the case of uk sector returns. Applied Financial Economics, 21(15):1079–1093, 2011

    2011

  16. [24]

    Variable selection for comparing high-dimensional time-series data.arXiv preprint arXiv:2412.06870, 2024

    Kensuke Mitsuzawa, Margherita Grossi, Stefano Bortoli, and Motonobu Kanagawa. Variable selection for comparing high-dimensional time-series data.arXiv preprint arXiv:2412.06870, 2024

    2024 arXiv

  17. [25]

    Variable selection in maximum mean discrepancy for interpretable distribution comparison.arXiv preprint arXiv:2311.01537, 2023

    Kensuke Mitsuzawa, Motonobu Kanagawa, Stefano Bortoli, Margherita Grossi, and Paolo Papotti. Variable selection in maximum mean discrepancy for interpretable distribution comparison.arXiv preprint arXiv:2311.01537, 2023

    2023

  18. [26]

    Oakland.Statistical Process Control

    John S. Oakland.Statistical Process Control. Routledge, London, 6 edition, 2007

    2007

  19. [27]

    Romano and Michael Wolf

    Joseph P. Romano and Michael Wolf. Exact and approximate stepdown methods for multiple hypothesis testing.Journal of the American Statistical Association, 100(469):94–108, 2005

    2005

  20. [28]

    Gordon J. Ross. Modelling financial volatility in the presence of abrupt changes.Physica A: Statistical Mechanics and its Applications, 392(2):350–360, 2013. 43

    2013

  21. [29]

    A permutation-free kernel two-sample test

    Shubhanshu Shekhar, Ilmun Kim, and Aaditya Ramdas. A permutation-free kernel two-sample test. Advances in Neural Information Processing Systems, 35:18168–18180, 2022

    2022

  22. [30]

    Hilbert space embeddings and metrics on probability measures.The Journal of Machine Learning Research, 11:1517–1561, 2010

    Bharath K Sriperumbudur, Arthur Gretton, Kenji Fukumizu, Bernhard Schölkopf, and Gert RG Lanckriet. Hilbert space embeddings and metrics on probability measures.The Journal of Machine Learning Research, 11:1517–1561, 2010

    2010

  23. [31]

    Valid and exact statistical inference for multi-dimensional multiple change-points by selective inference.arXiv preprint arXiv:2110.08989, 2021

    Ryota Sugiyama, Hiroki Toda, Vo Nguyen Le Duy, Yu Inatsu, and Ichiro Takeuchi. Valid and exact statistical inference for multi-dimensional multiple change-points by selective inference.arXiv preprint arXiv:2110.08989, 2021

    2021 arXiv

  24. [32]

    Industrial control anomaly detection based on distributed linear deep learning.Computers, Materials and Continua, 82(1):1129–1150, 2025

    Shijie Tang, Yong Ding, and Huiyong Wang. Industrial control anomaly detection based on distributed linear deep learning.Computers, Materials and Continua, 82(1):1129–1150, 2025

    2025

  25. [33]

    Selective review of offline change point detection methods.Signal Processing, 167:107299, 02 2020

    Charles Truong, Laurent Oudre, and Nicolas Vayatis. Selective review of offline change point detection methods.Signal Processing, 167:107299, 02 2020

    2020

  26. [34]

    Selective inference for change point detection in multi-dimensional sequences.arXiv preprint arXiv:1706.00514, 2017

    Yuta Umezu and Ichiro Takeuchi. Selective inference for change point detection in multi-dimensional sequences.arXiv preprint arXiv:1706.00514, 2017

    2017 arXiv

  27. [35]

    Paddy Farrington, Paul H

    Steffen Unkel, C. Paddy Farrington, Paul H. Garthwaite, Chris Robertson, and Nick Andrews. Statistical methods for the prospective detection of infectious disease outbreaks: a review.Journal of the Royal Statistical Society Series A: Statistics in Society, 175(1):49–82, 2012

    2012

  28. [36]

    Dey, and Yao Xie

    Jie Wang, Santanu S. Dey, and Yao Xie. Variable selection for kernel two-sample tests.arXiv preprint arXiv:2302.07415, 2023. 44

    2023 arXiv

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.