Search arXiv⌕ Search

arXiv · 2611.10094

Consistency considerations for rare-event importance sampling

Abstract

We consider the problem of estimating a small probability $p = P(X \in A)$ by importance sampling with auxiliary density $g$, where $X$ is drawn according to some density $f$. We establish a simple necessary and sufficient condition for the consistency of the importance sampling estimator in triangular schemes where the dimension, the densities $f$ and $g$ and the set $A$ may evolve. It allows to refine the results of Chatterjee and Diaconis [Ann. Appl. Probab. 28 (2018), no. 2, 1099-1135] by establishing the exact minimal sample size required for consistency. Moreover, it allows to study the regime $p \to 0$ which, despite being the relevant mathematical regime for rare-event estimations, has received very little attention so far. We perform an in-depth study of the hyperplane case $A = \{x: x^\top u \geq α\}$ with Gaussian densities, showing that in this case, the optimal importance sampling density considerably improves consistency compared to the standard Monte Carlo estimator. However, we conclude the paper with moment considerations which suggest that, whenever the Monte Carlo and optimal importance sampling estimators both converge, convergence may occur faster for the Monte Carlo estimator.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jason Beh, Emilien Flayac, Jerome Morio, Florian Simatos. 2026-11-06. Consistency considerations for rare-event importance sampling. https://arxiv.org/abs/2611.10094

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

Functional Extreme-PLS

An extreme dimension reduction method is proposed, extending the Extreme-PLS approach to the discretized functional framework, where the covariate lies in the infinite-dimensional Hilbert space ($L^2([0,1])$) but is partially observed on a dense time grid. The ideas are partly borrowed from both Partial Least-Squares (PLS) and Sliced Inverse Regression (SIR) techniques. Accordingly, the method relies on the projection of the covariate onto a subspace and maximizes the covariance between its projection and the response conditionally on an extreme event capturing tail information. The covariate and the heavy-tailed response are supposed to be linked through a nonlinear inverse single-index model, and the objective is to infer the index in this regression framework. A new family of estimators is proposed, and its asymptotic consistency with convergence rates under the model is established. Assuming mild conditions on the noise, most of the assumptions are stated in terms of regular variation, unlike those commonly imposed for SIR and single-index regression. In addition, the theoretical analysis is expanded to include a model-free consistency result in probability for the empirical tail-moments in a general separable Hilbert space. Finally, the results are illustrated through a finite-sample study with synthetic functional data and high-frequency financial data, highlighting the effectiveness of the dimension reduction in capturing tail dependence and supporting extreme risk management.

math.ST↗

Phase transition for conditional covariance matrices estimated by importance sampling, and implications for cross-entropy schemes in high dimension

Motivated by recent high-dimensional reliability algorithms, we study a random matrix model corresponding to the estimation by importance sampling of $Σ_A$, the covariance matrix of a Gaussian vector conditioned by a rare event $A$. For such models, we show that, in contrast to most random matrix models, the minimal sample size to ensure an accurate estimation of $Σ_A$ needs to grow superlinearly in the dimension. More precisely, if $n$ is the sample size and $d$ the dimension, we show that the estimation of $Σ_A$ is accurate if $n/d \gg L$ and not accurate if $n/d \ll L$ where $L$ is the maximum likelihood ratio. Moreover, we relate the asymptotic behavior of $L$ to $λ_1$, the smallest eigenvalue of the auxiliary covariance matrix, which suggests that $L$ (and thus the minimal sample size) will be larger for small values of $λ_1$. Numerical experiments confirm that this insight carries over to recent cross-entropy algorithms, with the best performing algorithms being also the ones with the largest smallest eigenvalue. Although backed up by our theoretical result, this result is surprising as these algorithms were not designed to act on the spectrum.

math.ST↗

A High Dimensional Wild Bootstrap Max-Test for Detecting the Presence of Significant Predictors

We construct a block bootstrap max-test for detecting the presence of significant predictors in a high dimensional setting, allowing for weakly dependent and heterogeneous (possibly non-stationary) data. The number of covariates to be screened may be large $p$ $>>$ $n$, and growing at an exponential rate, provided $\ln (p)$ $=$ $o(n^{a})$ for some $a$ $>$ $0$ that depends on memory decay and the growth of higher moments. We study the problem of correlation screening in a high dimensional marginal regression setting, assuming so-called \textit{physical dependence} in a time series setting. We entirely sidestep covariance matrix estimation and adaptive re-sampling by working with a max-statistic over the many computed parameters. Thus we do not need endogenous selection of the most relevant predictor index yielding non-uniform asymptotics, nor do we need a post-estimation Bonferroni correction. The sequence of non-standard approximating distributions arising from the maximum of an increasing number of estimators is easily approximated by a multiplier (wild) block bootstrap. The max-test controls for size well, performs well against various deviations from the null, including very slight deviations with a weak or sparse signal. A numerical experiment is performed and an empirical example with the VIX volatility index is provided.

math.ST↗