Convergence of the empirical two-sample -statistics with -mixing data

7

and since [ntu] 6 [ntu0−1] + 1 for n large enough, we get

α (j) ;

(2.17)

n Cov

^Yn,i^,

Yn,i

6 n

[ntu0−1]

+∞

1

i∈In,u

i^0∈In,u0

K

i+1

i=1

j=[ntu0−1]

−

X

X

X

X

= [nt

u0

−1

]

−

i gives

doing the change k

^6 n

α (j)

(2.18)