On weakly bounded empirical processes

Abstract: Let

F

be a class of functions on a probability space

(O m e g a, m u)

and let

X_{1}, ..., X_{k}

be independent random variables distributed according to

m u

. We establish high probability tail estimates of the form

s u p_{f i n F} | i : | f (X_{i}) | g e q t

using a natural parameter associated with

F

. We use this result to analyze weakly bounded empirical processes indexed by

F

and processes of the form

Z_{f} = | k^{- 1} s u m_{i = 1}^{k} | f |^{p} (X_{i}) - E | f |^{p} |

for

p > 1

. We also present some geometric applications of this approach, based on properties of the random operator

G a m m a = k^{- 1 / 2} s u m_{i = 1}^{k} i n r X_{i}, c d o t e_{i}

, where the

(X_{i})_{i = 1}^{k}

are sampled according to an isotropic, log-concave measure on

R^{n}

.

Recommendations

Cites Work

Cited In (13)

This page was built for publication: On weakly bounded empirical processes