Projections of random vectors via the generic chaining
Author:
- Elizabeth MeckesEMElizabeth Meckes
Preprint
Abstract
Many authors have studied the phenomenon of typically Gaussian marginals of high-dimensional random vectors. In earlier work, the author used entropy methods and Stein's method to show that this phenomenon persists for $k$-dimensional marginals of $d$-dimensional distributions, if $k=o(\sqrt{\log(d)})$. In this paper it is shown that the phenomenon persists if $k=o(\frac{\log(d)}{\log(\log(d))})$, and that this estimate is best possible. The proof uses the generic chaining bound of M.\ Talagrand and his application of it to stochastic processes indexed by balls in $p$-convex spaces.
Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- Approximation of Projections of Random VectorsJournal of Theoretical Probability, 2010
- Power-law estimates for the central limit theorem for convex setsJournal of Functional Analysis, 2007
- A central limit theorem for convex setsInventiones Mathematicae, 2007
- On concentration of distributions of random weighted sumsThe Annals of Probability, 2003
- Asymptotics of Graphical Projection PursuitThe Annals of Statistics, 1984