Tail bounds for empirically standardized sums
From MaRDI portal
Publication:2137797
Abstract: Exponential tail bounds for sums play an important role in statistics, but the example of the -statistic shows that the exponential tail decay may be lost when population parameters need to be estimated from the data. However, it turns out that if Studentizing is accompanied by estimating the location parameter in a suitable way, then the -statistic regains the exponential tail behavior. Motivated by this example, the paper analyzes other ways of empirically standardizing sums and establishes tail bounds that are sub-Gaussian or even closer to normal for the following settings: Standardization with Studentized contrasts for normal observations, standardization with the log likelihood ratio statistic for observations from an exponential family, and standardization via self-normalization for observations from a symmetric distribution with unknown center of symmetry. The latter standardization gives rise to a novel scan statistic for heteroscedastic data whose asymptotic power is analyzed in the case where the observations have a log-concave distribution.
Recommendations
Cites work
- A Cramér type large deviation result for Student's \(t\)-statistic
- A spatial scan statistic
- A tight Gaussian bound for weighted sums of Rademacher random variables
- An asymptotically Gaussian bound on the Rademacher tails
- Bounds on tail probabilities for negative binomial distributions.
- Bump detection in heterogeneous Gaussian regression
- Concentration inequalities. A nonasymptotic theory of independence
- Cramér type moderate deviation theorems for self-normalized processes
- High-dimensional probability. An introduction with applications in data science
- High-dimensional statistics. A non-asymptotic viewpoint
- scientific article; zbMATH DE number 3978129 (Why is no real title available?)
- scientific article; zbMATH DE number 44577 (Why is no real title available?)
- Inference and modeling with log-concave distributions
- Log-concavity and strong log-concavity: a review
- Multidimensional multiscale scanning in exponential families: limit theory and statistical consequences
- Multiscale change point inference. With discussion and authors' reply
- Optimal and fast detection of spatial clusters with scan statistics
- Optimal detection of a jump in the intensity of a Poisson process or in a density with likelihood ratio statistics
- Recent progress in log-concave density estimation
- Self-normalization: taming a wild population in a heavy-tailed world
- Self-normalized Cramér-type large deviations for independent random variables.
- Self-Normalized Processes
- Student's t-Test Under Symmetry Conditions
- Tests for change-points with epidemic alternatives
- Toward the best constant factor for the Rademacher-Gaussian tail comparison
- Weak convergence and empirical processes. With applications to statistics
- When is the Student \(t\)-statistic asymptotically standard normal?
Cited in
(2)
This page was built for publication: Tail bounds for empirically standardized sums
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2137797)
