arXiv subjects

Holger Dette

Publications and source records attributed to Holger Dette.

At least 37 records · Page 2Linked to original sources

arXiv2026

Inference for Multiple Change-points in Piecewise Locally Stationary Time Series

Change-point detection and locally stationary time series modeling are two major approaches for the analysis of non-stationary data. The former aims to identify stationary phases by detecting abrupt changes in the dynamics of a time series model, while the latter employs (locally) time-varying models to describe smooth changes in dependence structure of a time series. However, in some applications, abrupt and smooth changes can co-exist, and neither of the two approaches alone can model the data adequately. In this paper, we propose a novel likelihood-based procedure for the inference of multiple change-points in locally stationary time series. In contrast to traditional change-point analysis where an abrupt change occurs in a real-valued parameter, a change in locally stationary time series occurs in a parameter curve, and can be classified as a jump or a kink depending on whether the curve is discontinuous or not. We show that the proposed method can consistently estimate the number, locations, and the types of change-points. Two different asymptotic distributions corresponding respectively to jump and kink estimators are also established. Extensive simulation studies and a real data application to financial time series are provided.

stat.ME↗

arXiv2025

Inference for Forecasting Accuracy: Pooled versus Individual Estimators in High-dimensional Panel Data

Panels with large time $(T)$ and cross-sectional $(N)$ dimensions are a key data structure in social sciences and other fields. A central question in panel data analysis is whether to pool data across individuals or to estimate separate models. Pooled estimators typically have lower variance but may suffer from bias, creating a fundamental trade-off for optimal estimation. We develop a new inference method to compare the forecasting performance of pooled and individual estimators. Specifically, we propose a confidence interval for the difference between their forecasting errors and establish its asymptotic validity. Our theory allows for complex temporal and cross-sectional dependence in the model errors and covers scenarios where $N$ can be much larger than $T$-including the independent case under the classical condition $N/T^2 \to 0$. The finite-sample properties of the proposed method are examined in an extensive simulation study.

stat.ME↗

arXiv2025

Testing for equivalence of pre-trends in Difference-in-Differences estimation

The plausibility of the ``parallel trends assumption'' in Difference-in-Differences estimation is usually assessed by a test of the null hypothesis that the difference between the average outcomes of both groups is constant over time before the treatment. However, failure to reject the null hypothesis does not imply the absence of differences in time trends between both groups. We provide equivalence tests that allow researchers to find evidence in favor of the parallel trends assumption and thus increase the credibility of their treatment effect estimates. While we motivate our tests in the standard two-way fixed effects model, we discuss simple extensions to settings in which treatment adoption is staggered over time.

econ.EM↗

arXiv2025

Convergence of covariance and spectral density estimates for high-dimensional functional time series

Second-order characteristics including covariance and spectral density functions are fundamentally important for both statistical applications and theoretical analysis in functional time series. In the high-dimensional setting where the number of functional variables is large relative to the length of functional time series, non-asymptotic theory for covariance function estimation has been developed for Gaussian and sub-Gaussian functional linear processes. However, corresponding non-asymptotic results for high-dimensional non-Gaussian and nonlinear functional time series, as well as for spectral density function estimation, are largely unexplored. In this paper, we introduce novel functional dependence measures, based on which we establish systematic non-asymptotic concentration bounds for estimates of (auto)covariance and spectral density functions in high-dimensional and non-Gaussian settings. We then illustrate the usefulness of our convergence results through two applications to dynamic functional principal component analysis and sparse spectral density function estimation. To handle the practical scenario where curves are discretely observed with errors, we further develop convergence rates of the corresponding estimates obtained via a nonparametric smoothing method. Finally, extensive simulation studies are conducted to corroborate our theoretical findings.

math.ST↗

arXiv2025

Detecting practically significant dependencies in metric spaces via distance correlations

We take a different look at the problem of testing the independence of two metric-space-valued random variables using the distance correlation. Instead of testing if the distance correlation vanishes exactly, we are interested in the hypothesis that it does not exceed a certain threshold. Our testing problem is motivated by the observation that in many cases it is more reasonable to test for a practically significant dependency since it is rare that a hypothesis of perfect independence is exactly satisfied. This point of view also reflects statistical practice, where one often classifies the strength of the association in categories such as `small', `medium' and `large' and the precise definitions depend on the specific application. To address these problems we develop a pivotal test for the hypothesis that the distance correlation between two random variables does not exceed a pre-specified threshold $Δ$. We also determine a minimum value $\hat Δ_α$ from the data such that the hypothesis is rejected for all $Δ\leq \hat Δ_α$ at controlled type I error $α$. This quantity can be interpreted as a measure of evidence against the hypothesis that the distance correlation is less or equal than $Δ$. The new test is applicable to processes taking values in separable metric spaces of strong negative type, covering Euclidean as well as functional data. We do not assume independent observations, and instead prove our results for absolutely regular sample generating processes, which includes many time series such as ARMA and GARCH models. Our approach is based on a new functional limit theorem for the sequential distance correlation process, and can also be used to construct confidence intervals for the distance correlation without the need for resampling.

math.ST↗

arXiv2025

Multiscale Change Point Detection for Functional Time Series

We study the problem of detecting and localizing multiple changes in the mean parameter of a Banach space-valued time series. The goal is to construct a collection of narrow confidence intervals, each containing at least one (or exactly one) change, with globally controlled error probability. Our approach relies on a new class of weighted scan statistics, called Hölder-type statistics, which allow a smooth trade-off between efficiency (enabling the detection of closely spaced, small changes) and robustness (against heavier tails and stronger dependence). For Gaussian noise, maximum weighting can be applied, leading to a generalization of optimality results known for scalar, independent data. Even for scalar time series, our approach is advantageous, as it accommodates broad classes of dependency structures and non-stationarity. Its primary advantage, however, lies in its applicability to functional time series, where few methods exist and established procedures impose strong restrictions on the spacing and magnitude of changes. We obtain general results by employing new Gaussian approximations for the partial sum process in Hölder spaces. As an application of our general theory, we consider the detection of distributional changes in a data panel. The finite-sample properties and applications to financial datasets further highlight the merits of our method.

math.ST↗

arXiv2025

Monitoring Violations of Differential Privacy over Time

Auditing differential privacy has emerged as an important area of research that supports the design of privacy-preserving mechanisms. Privacy audits help to obtain empirical estimates of the privacy parameter, to expose flawed implementations of algorithms and to compare practical with theoretical privacy guarantees. In this work, we investigate an unexplored facet of privacy auditing: the sustained auditing of a mechanism that can go through changes during its development or deployment. Monitoring the privacy of algorithms over time comes with specific challenges. Running state-of-the-art (static) auditors repeatedly requires excessive sampling efforts, while the reliability of such methods deteriorates over time without proper adjustments. To overcome these obstacles, we present a new monitoring procedure that extracts information from the entire deployment history of the algorithm. This allows us to reduce sampling efforts, while sustaining reliable outcomes of our auditor. We derive formal guarantees with regard to the soundness of our methods and evaluate their performance for important mechanisms from the literature. Our theoretical findings and experiments demonstrate the efficacy of our approach.

cs.CR↗

arXiv2025

Monitoring Time Series for Relevant Changes

We consider the problem of sequentially testing for changes in the mean parameter of a time series, compared to a benchmark period. Most tests in the literature focus on the null hypothesis of a constant mean versus the alternative of a single change at an unknown time. Yet in many applications it is unrealistic that no change occurs at all, or that after one change the time series remains stationary forever. We introduce a new setup, modeling the sequence of means as a piecewise constant function with arbitrarily many changes. Instead of testing for a change, we ask whether the evolving sequence of means, say $(μ_n)_{n \geq 1}$, stays within a narrow corridor around its initial value, that is, $μ_n \in [μ_1-Δ, μ_1+Δ]$ for all $n \ge 1$. Combining elements from multiple change point detection with a Hölder-type monitoring procedure, we develop a new online monitoring tool. A key challenge in both construction and proof of validity is that the risk of committing a type-I error after any time $n$ fundamentally depends on the unknown future of the time series. Simulations support our theoretical results and we present two real-world applications: (1) healthcare monitoring, with a focus on blood glucose tracking, and (2) political consensus analysis via citizen opinion polls.

stat.ME↗

arXiv2025

Practically significant change points in high dimension -- measuring signal strength pro active component

We consider the change point testing problem for high-dimensional time series. Unlike conventional approaches, where one tests whether the difference $δ$ of the mean vectors before and after the change point is equal to zero, we argue that the consideration of the null hypothesis $H_0:\|δ\|\leΔ$, for some norm $\|\cdot\|$ and a threshold $Δ>0$, is better suited. By the formulation of the null hypothesis as a composite hypothesis, the change point testing problem becomes significantly more challenging. We develop pivotal inference for testing hypotheses of this type in the setting of high-dimensional time series, first, measuring deviations from the null vector by the $\ell_2$-norm $\|\cdot\|_2$ normalized by the dimension. Second, by measuring deviations using a sparsity adjusted $\ell_2$-"norm" $\|\cdot \|_2/\sqrt{\|\cdot\|_0} $, where $\|\cdot\|_0$ denotes the $\ell_0$-"norm," we propose a pivotal test procedure which intrinsically adapts to sparse alternatives in a data-driven way by pivotally estimating the set of nonzero entries of the vector $δ$. To establish the statistical validity of our approach, we derive tail bounds of certain classes of distributions that frequently appear as limiting distributions of self-normalized statistics. As a theoretical foundation for all results, we develop a general weak invariance principle for the partial sum process $X_1^\topξ+\cdots +X_{\lfloorλn\rfloor}^\topξ$ for a time series $(X_j)_{j\in\mathbb{Z}}$ and a contrast vector $ξ\in\mathbb{R}^p$ under increasing dimension $p$, which is of independent interest. Finally, we investigate the finite sample properties of the tests by means of a simulation study and illustrate its application in a data example.

math.ST↗

arXiv2025

General-Purpose $f$-DP Estimation and Auditing in a Black-Box Setting

In this paper we propose new methods to statistically assess $f$-Differential Privacy ($f$-DP), a recent refinement of differential privacy (DP) that remedies certain weaknesses of standard DP (including tightness under algorithmic composition). A challenge when deploying differentially private mechanisms is that DP is hard to validate, especially in the black-box setting. This has led to numerous empirical methods for auditing standard DP, while $f$-DP remains less explored. We introduce new black-box methods for $f$-DP that, unlike existing approaches for this privacy notion, do not require prior knowledge of the investigated algorithm. Our procedure yields a complete estimate of the $f$-DP trade-off curve, with theoretical guarantees of convergence. Additionally, we propose an efficient auditing method that empirically detects $f$-DP violations with statistical certainty, merging techniques from non-parametric estimation and optimal classification theory. Through experiments on a range of DP mechanisms, we demonstrate the effectiveness of our estimation and auditing procedures.

cs.CR↗

arXiv2025

Practically significant differences between conditional distribution functions

In the framework of semiparametric distribution regression, we consider the problem of comparing the conditional distribution functions corresponding to two samples. In contrast to testing for exact equality, we are interested in the (null) hypothesis that the $L^2$ distance between the conditional distribution functions does not exceed a certain threshold in absolute value. The consideration of these hypotheses is motivated by the observation that in applications, it is rare, and perhaps impossible, that a null hypothesis of exact equality is satisfied and that the real question of interest is to detect a practically significant deviation between the two conditional distribution functions. The consideration of a composite null hypothesis makes the testing problem challenging, and in this paper we develop a pivotal test for such hypotheses. Our approach is based on self-normalization and therefore requires neither the estimation of (complicated) variances nor bootstrap approximations. We derive the asymptotic limit distribution of the (appropriately normalized) test statistic and show consistency under local alternatives. A simulation study and an application to German SOEP data reveal the usefulness of the method.

econ.EM↗

arXiv2025

A New Two-Sample Test for Covariance Matrices in High Dimensions: U-Statistics Meet Leading Eigenvalues

We propose a two-sample test for covariance matrices in the high-dimensional regime, where the dimension diverges proportionally to the sample size. Our hybrid test combines a Frobenius-norm-based statistic as considered in Li and Chen (2012) with the leading eigenvalue approach proposed in Zhang et al. (2022), making it sensitive to both dense and sparse alternatives. The two statistics are combined via Fisher's method, leveraging our key theoretical result: a joint central limit theorem showing the asymptotic independence of the leading eigenvalues of the sample covariance matrix and an estimator of the Frobenius norm of the difference of the two population covariance matrices, under suitable signal conditions. The level of the test can be controlled asymptotically, and we show consistency against certain types of both sparse and dense alternatives. A comprehensive numerical study confirms the favorable performance of our method compared to existing approaches.

math.ST↗

arXiv2025

Gaussian Differential Private Bootstrap by Subsampling

Bootstrap is a common tool for quantifying uncertainty in data analysis. However, besides additional computational costs in the application of the bootstrap on massive data, a challenging problem in bootstrap based inference under Differential Privacy consists in the fact that it requires repeated access to the data. As a consequence, bootstrap based differentially private inference requires a significant increase of the privacy budget, which on the other hand comes with a substantial loss in statistical accuracy. A potential solution to reconcile the conflicting goals of statistical accuracy and privacy is to analyze the data under parametric model assumptions and in the last decade, several parametric bootstrap methods for inference under privacy have been investigated. However, uncertainty quantification by parametric bootstrap is only valid if the the quantities of interest can be identified as the parameters of a statistical model and the imposed model assumptions are (at least approximately) satisfied. An alternative to parametric methods is the empirical bootstrap that is a widely used tool for non-parametric inference and well studied in the non-private regime. However, under privacy, less insight is available. In this paper, we propose a private empirical $m$ out of $n$ bootstrap and validate its consistency and privacy guarantees under Gaussian Differential Privacy. Compared to the the private $n$ out of $n$ bootstrap, our approach has several advantages. First, it comes with less computational costs, in particular for massive data. Second, the proposed procedure needs less additional noise in the bootstrap iterations, which leads to an improved statistical accuracy while asymptotically guaranteeing the same level of privacy. Third, we demonstrate much better finite sample properties compared to the currently available procedures.

stat.ML↗

arXiv2025

SILENT: A New Lens on Statistics in Software Timing Side Channels

Cryptographic research takes software timing side channels seriously. Approaches to mitigate them include constant-time coding and techniques to enforce such practices. However, recent attacks like Meltdown [42], Spectre [37], and Hertzbleed [70] have challenged our understanding of what it means for code to execute in constant time on modern CPUs. To ensure that assumptions on the underlying hardware are correct and to create a complete feedback loop, developers should also perform \emph{timing measurements} as a final validation step to ensure the absence of exploitable side channels. Unfortunately, as highlighted by a recent study by Jancar et al. [30], developers often avoid measurements due to the perceived unreliability of the statistical analysis and its guarantees. In this work, we combat the view that statistical techniques only provide weak guarantees by introducing a new algorithm for the analysis of timing measurements with strong, formal statistical guarantees, giving developers a reliable analysis tool. Specifically, our algorithm (1) is non-parametric, making minimal assumptions about the underlying distribution and thus overcoming limitations of classical tests like the t-test, (2) handles unknown data dependencies in measurements, (3) can estimate in advance how many samples are needed to detect a leak of a given size, and (4) allows the definition of a negligible leak threshold $Δ$, ensuring that acceptable non-exploitable leaks do not trigger false positives, without compromising statistical soundness. We demonstrate the necessity, effectiveness, and benefits of our approach on both synthetic benchmarks and real-world applications.

cs.CR↗

arXiv2025

Multiscale detection of practically significant changes in a gradually varying time series

In many change point problems it is reasonable to assume that compared to a benchmark at a given time point $t_0$ the properties of the observed stochastic process change gradually over time for $t >t_0$. Often, these gradual changes are not of interest as long as they are small (nonrelevant), but one is interested in the question if the deviations are practically significant in the sense that the deviation of the process compared to the time $t_0$ (measured by an appropriate metric) exceeds a given threshold, which is of practical significance (relevant change). In this paper we develop novel and powerful change point analysis for detecting such deviations in a sequence of gradually varying means, which is compared with the average mean from a previous time period. Current approaches to this problem suffer from low power, rely on the selection of smoothing parameters and require a rather regular (smooth) development for the means. We develop a multiscale procedure that alleviates all these issues, validate it theoretically and demonstrate its good finite sample performance on both synthetic and real data.

stat.ME↗

arXiv2025

Exponential Inequalities for Some Mixing Processes and Dynamic Systems

Many important dynamic systems, time series models or even algorithms exhibit non-strong mixing properties. In this paper, we introduce the general concept of $\mathcal{C}_{p,\mathcal{F}}$-mixing to cover such cases, where assumptions on the dependence structure become stronger with increasing $p\in [1, \infty].$ We derive a series of sharp exponential-type (or Bernstein-type) inequalities under this dependence concept for $p=1$ and $p=\infty$. More specifically, $\mathcal{C}_{\infty,\mathcal{F}}$-mixing is equal to the widely discussed $\mathcal{C}$-mixing \citep{maume2006exponential}, and we prove a refinement of an Berntsein-type inequality in \cite{hang2017bernstein} for $\mathcal{C}$-mixing processes under more general assumptions. As there exist many stochastic processes and dynamic systems, which are not $\mathcal{C}$ (or $\mathcal{C}_{\infty,\mathcal{F}}$)-mixing, we derive Bernstein-type inequalities for $\mathcal{C}_{1,\mathcal{F}}$-mixing processes as well and we use this result to investigate the convergence rates of plug-in-type estimators of the local conditional mode set for vector-valued output, in particular in situations where the density is less smooth.

math.ST↗

arXiv2025

Sequential Outlier Detection in Non-Stationary Time Series

A novel method for sequential outlier detection in non-stationary time series is proposed. The method tests the null hypothesis of ``no outlier'' at each time point, addressing the multiple testing problem by bounding the error probability of successive tests, using extreme value theory. The asymptotic properties of the test statistic are studied under the null hypothesis and alternative. The finite sample properties of the new detection scheme are investigated by means of a simulation study, and the method is compared with alternative procedures which have recently been proposed in the statistics and machine learning literature.

math.ST↗

arXiv2025

Uniform confidence bands for joint angles across different fatigue phases

We develop uniform confidence bands for the mean function of stationary time series as a post-hoc analysis of multiple change point detection in functional time series. In particular, the methodology in this work provides bands for those segments where the jump size exceeds a certain threshold $Δ$. In \cite{bastian2024multiplechangepointdetection} such exceedences of $Δ$ were related to fatigue states of a running athlete. The extension to confidence bands stems from an interest in understanding the range of motion (ROM) of lower-extremity joints of running athletes under fatiguing conditions. From a biomechanical perspective, ROM serves as a proxy for joint flexibility under varying fatigue states, offering individualized insights into potentially problematic movement patterns. The new methodology provides a valuable tool for understanding the dynamic behavior of joint motion and its relationship to fatigue.

stat.ME↗