On normal approximations to U-statistics

Asymptotic properties of nonparametric inference (62G20) Asymptotic distribution theory in statistics (62E20) Central limit and other weak theorems (60F05)

Abstract: Let

X_{1}, ..., X_{n}

be i.i.d. random observations. Let

m a t h b b S = m a t h b b L + m a t h b b T

be a

U

-statistic of order

k g e 2

where

m a t h b b L

is a linear statistic having asymptotic normal distribution, and

m a t h b b T

is a stochastically smaller statistic. We show that the rate of convergence to normality for

m a t h b b S

can be simply expressed as the rate of convergence to normality for the linear part

m a t h b b L

plus a correction term,

(o p e r a t o r n a m e v a r m a t h b b T) l n^{2} (o p e r a t o r n a m e v a r m a t h b b T)

, under the condition

m a t h b b E m a t h b b T^{2} < i n f t y

. An optimal bound without this

l o g

factor is obtained under a lower moment assumption

m a t h b b E | m a t h b b T |^{a l p h a} < i n f t y

for

a l p h a < 2

. Some other related results are also obtained in the paper. Our results extend, refine and yield a number of related-known results in the literature.

Recommendations

Cites work

Cited in

(22)

This page was built for publication: On normal approximations to \(U\)-statistics

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q971937)