{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Lecture 15-2007

# Lecture 15-2007 - Lecture XIV Let represent the entire...

This preview shows pages 1–10. Sign up to view the full content.

Lecture XIV

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

View Full Document
Let  ϖ  represent the entire random sequence  { Z t }.  As discussed last time, our interest  typically centers around the averages of this  sequence: ( 29 = = n t t n Z n b 1 1 ω
Definition 2.9: Let { b n ( ϖ )} be a sequence of  real-valued random variables.  We say that  b n ( ϖ ) converges  almost surely  to  b , written if and only if there exists a real number  b  such  that ( 29 b b s a n → . . ω ( 29 [ ] 1 : = b b P n ω ω

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

View Full Document
The probability measure  P  describes the  distribution of  ϖ  and determines the joint  distribution function for the entire sequence  { Z t }. Other common terminology is that  b n ( ϖ converges to b  with probability 1  (w.p.1) or  that  b n ( ϖ is strongly  consistent for  b .
Example 2.10: Let  where { Z t } is a sequence of independently and  identically distributed (i.i.d.) random variables  with  E ( Z t )= μ < .  Then by the Komolgorov strong law of large  numbers (Theorem 3.1). = = n t t n Z n Z 1 1 μ → . . s a n Z

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

View Full Document
Proposition 2.11: Given  g R k R l  ( k , l < ) and  any sequence { b n } such that  where  b n  and  b  are  k  x 1 vectors, if  g  is  continuous at  b , then b b s a n → . . ( 29 ( 29 b g b g s a n → . .
Theorem 2.12: Suppose y = X β 0 + ε ; X ε / n   a.s.  0; X X / a.s. M , finite and positive definite. Then  β n  exists  a.s.  for all  n  sufficiently large,  and  β n a.s. β 0 .

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

View Full Document
Proof: Since  X X / n   a.s. M , it follows from  Proposition 2.11 that det( X X / n a.s. det( M ).   Because  M  is positive definite by (iii),  det( M )>0.  It follows that det( X X / n )>0 a.s. for  all  n  sufficiently large, so ( X X / n ) -1  exists a.s.  for all  n  sufficiently large.  Hence ( 29 n y X n X X n ' ' ˆ 1 - β
In addition, It follows from Proposition 2.11 that ( 29 n X n X X n ε β β ' ' ˆ 1 0 - + = 0 1 0 . . 0 0

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

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}