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.

201<br />

• Likelihood ratio test. Put<br />

∆ = <br />

³ˆ ´<br />

− ( 0 )=log <br />

³ˆ ´<br />

( 0 ) <br />

Under this is expected to be large.<br />

Lemma: Under 2∆ = 2 + (1).<br />

By this, (recall sample midterm exam question, or<br />

Lemma 7.7.1) the tests based on and 2∆ are<br />

asymptotically equivalent. That based on 2∆ <br />

rejects for<br />

2∆ 2 1 () <br />

(For one-sided alternatives the ‘signed likelihood<br />

ratio’ √ 2∆ · (ˆ − 0 ) can be used.) An<br />

asymptotic 1 − confidence region is<br />

= n 0 | 2∆ ( 0 ) ≤ 2 1 ()o

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

Saved successfully!

Ooh no, something went wrong!