PDF of Lecture Notes - School of Mathematical Sciences
PDF of Lecture Notes - School of Mathematical Sciences
PDF of Lecture Notes - School of Mathematical Sciences
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
1. DISTRIBUTION THEORY<br />
Hence we have shown M Zn (t) = (1 + a n ) n , where<br />
lim na n = t 2 /2, so by Lemma 1.11.1,<br />
n→∞<br />
lim M Z n<br />
(t) = e t2 /2 for each fixed t.<br />
n→∞<br />
Remarks<br />
1. The Central Limit Theorem can be stated equivalently for ¯X n ,<br />
( ) ¯Xn − µ<br />
i.e., L<br />
σ/ √ → N(0, 1)<br />
n<br />
(<br />
just note that ¯X n − µ<br />
σ/ √ n = S n − nµ<br />
σ √ n<br />
2. The Central Limit Theorem holds under conditions more general than those given<br />
above. In particular, with suitable assumptions,<br />
(i) M X (t) need not exist.<br />
(ii) X 1 , X 2 , . . . need not be i.i.d..<br />
3. Theorems 1.11.1 and 1.11.2 are concerned with the asymptotic behaviour <strong>of</strong> ¯X n .<br />
)<br />
.<br />
Theorem 1.11.1 states ¯X n → µ in prob as n → ∞.<br />
Theorem 1.11.2 states ¯X n − µ<br />
σ/ √ n −→ D<br />
N(0, 1) as n → ∞.<br />
These results are not contradictory because Var( ¯X n ) → 0, but the Central Limit<br />
Theorem is concerned with ¯X n − E( ¯X n )<br />
√<br />
Var( ¯Xn ) . 76