29.07.2014 Views

Asymptotic Methods in Statistical Inference - Statistics Centre

Asymptotic Methods in Statistical Inference - Statistics Centre

Asymptotic Methods in Statistical Inference - Statistics Centre

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

162<br />

• Extend (i) to<br />

⎛√ ⎞<br />

<br />

³ˆ − ( )´<br />

( )= ⎝ ≤ ⎠ <br />

( )<br />

(18.1)<br />

for some scale functional ( ). Now ( )isestimated<br />

by ( ˆ ) and the latter is approximated<br />

by ∗ <br />

. The total error is<br />

where<br />

∗ − ( )= ³ ∗ − ( ˆ )´ + <br />

<br />

<br />

= ( ˆ ) − ( )<br />

is called the ‘bootstrap error’ . Typically ∗ −<br />

( ˆ ) can be made arbitrarily small by tak<strong>in</strong>g<br />

enough resamples. We say that ‘the bootstrap<br />

<br />

works’ if → 0. This <strong>in</strong>volves more than the<br />

cont<strong>in</strong>uity of the functional (·), s<strong>in</strong>ce itself<br />

varies with . If the bootstrap works, then<br />

∗ − ( ) → 0as →∞<br />

Our analysis of will proceed by subtract<strong>in</strong>g,<br />

from each of its terms, the limit (assumed to exist)<br />

( )of ( ); <strong>in</strong> the simplest cases this limit<br />

does not depend on .

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!