83 the multivariate clt implies that var x n 12 x n

Info iconThis preview shows pages 83–88. Sign up to view the full content.

View Full Document Right Arrow Icon
83
Background image of page 83

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
The multivariate CLT implies that Var X ̄ n  1/2 X ̄ n 1/2 n X ̄ n d Normal 0 , I k Note that the standardized random vector Z n 1/2 n X ̄ n has E Z n 0 and Var Z n I k . We can instead write n X ̄ n d Normal 0 , 84
Background image of page 84
8 . Some Additional Useful Limit Results Equipped with the multivariate CLT and the algebra of convergence in probability and convergence in distribution, we can state several limit results that have application for asymptotic analysis. (i) Suppose A n is a seqence of random r k matrices with A n p A , and W n is a sequence of k 1 random vectors with W n d W . Then A n W n d AW 85
Background image of page 85

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
To show this, write A n W n AW n A n A W n AW n o p 1 O p 1 AW n o p 1 and use AW n d AW along with the asymptotic equivalence lemma. If W ~ Normal 0 , then A n W n d Normal 0 , A A This usually applies when W n n 1/2 X ̄ n and the CLT is in force, and then A n n 1/2 X ̄ n  d Normal 0 , A A 86
Background image of page 86
(ii) Under the same setting above with W Normal 0 , , nonsingular, and rank A r , W n A n A A 1 A n W n d r 2 This follows because the quadratic function is continuous in A n W n and W A A A 1 AW r 2 . (iii) In the same setup as part (ii), let n be a sequence of symmetric, PSD matrices with n p . Then W n A n A n n A n 1 A n W n d r 2 87
Background image of page 87

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
This is a little harder to show, but we use the asymptotic equivalance lemma by showing that W n A n A n n A n 1 A n W n W n A n A A 1 A n W n o p 1 . To see this, note that W n A n A n n A n 1 A n W n W n A n A A 1 A n W n W n A n  A n n A n 1 A A 1 A n W n O p 1  o p 1 O p 1 because A n n A n 1 A A 1 p 0 . Now we can apply the asymptotic equivalence lemma. 88
Background image of page 88
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

Page83 / 88

83 The multivariate CLT implies that Var X n 12 X n 12 n X...

This preview shows document pages 83 - 88. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online