05.12.2012 Views

Student Notes To Accompany MS4214: STATISTICAL INFERENCE

Student Notes To Accompany MS4214: STATISTICAL INFERENCE

Student Notes To Accompany MS4214: STATISTICAL INFERENCE

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

hence<br />

E( √ � ∞ √ 1<br />

Z) = z<br />

2Γ(1) e−z/2dz = 23/2Γ(3/2) 2Γ(1) = � π/2 ,<br />

0<br />

hence ˜σ = 2ˆσ/ √ π is unbiased for σ. If we let d = (n − 1) we similarly find for<br />

general n that<br />

E( √ � ∞ √<br />

Z) = z<br />

so<br />

0<br />

zd/2−1 2d/2Γ(d/2) e−z/2dz = 2(d+1)/2Γ((d + 1)/2)<br />

2d/2Γ(d/2) c =<br />

makes cS an unbiased estimator of σ.<br />

√ dΓ(d/2)<br />

√ 2Γ([d + 1]/2)<br />

7. First of all c(θ) −1 = � ∞<br />

0 x2 e −θx dx = Γ(3)/θ 3 = 2/θ 3 . Next, we get E( ˜ θ) =<br />

E(2/X) = θ 3 � ∞<br />

0 xe−θx dx = θ 3 θ −2 Γ(2) = θ and so ˜ θ is an unbiased estimator of θ.<br />

<strong>To</strong> get the variance we use E( ˜ θ 2 ) = 2θ 3 � ∞<br />

0 e−θx dx = 2θ 2 so Var( ˜ θ) = 2θ 2 −θ 2 = θ 2 .<br />

<strong>To</strong> find the Fisher information for θ, we first calculate ∂ ln f(x|θ) = −x + 3/θ,<br />

∂θ<br />

and then calculate − ∂2<br />

∂θ2 ln f(x|θ) = 3/θ2 . Thus eff( ˜ θ) = θ2 /(3θ2 ) = 1<br />

3 .<br />

E(ˆµ) = θ3<br />

6<br />

� ∞<br />

0 x3e−θxdx = θ3Γ(4) 6θ4 = 1/θ = µ so ˆµ is an unbiased estimator of µ.<br />

Similarly E(ˆµ 2 ) = θ3<br />

18<br />

� ∞<br />

0 x4 e −θx dx = θ3 Γ(5)<br />

θ 5 = 4<br />

3θ 2 so Var(ˆµ) = 1<br />

3θ 2 . <strong>To</strong> calculate<br />

the CRLB we find for g(θ) = 1/θ that g ′ (θ) = −θ −2 so the lower bound is<br />

I(θ) −1 {g ′ (θ) 2 } = θ2<br />

3 θ−4 = 1<br />

= Var(ˆµ).<br />

3θ2 8. The formal definition of a sufficient statistic (S) is that the conditional distribu-<br />

tion of the sample X = (X1, X2, . . . , Xn) given the value of S, that is knowing<br />

that S = s, does not depend on θ.<br />

If f(x|θ) is the joint distribution of the sample X, the statistic S is sufficient for<br />

θ if and only if there exists function g(s|θ) and h(x) such that, for all sample<br />

points {xi} and all θ, the density factorises f(x|θ) = g(s|θ)h(x).<br />

f(x|θ) =<br />

n�<br />

x<br />

i=1<br />

θ−1<br />

i e−xi {Γ(θ)} n<br />

= exp[(θ − 1) � ln(xi)]<br />

{Γ(θ)} n<br />

× e − � xi<br />

↑ ↑<br />

�� �<br />

g ln(xi)|θ h(x)<br />

Thus, by the factorisation theorem, S = n�<br />

ln(Xi) is sufficient for θ.<br />

47<br />

i=1<br />

,

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

Saved successfully!

Ooh no, something went wrong!