ANSWERS TO EXERCISES FROM CHAPTER 3,

STA-2002

3.1
(
1
fX (x) = N1 −N0 +1 , x = N0 , N0 + 1, . . . , N1 .
0, otherwise
N0 +N1
E(X) = 2
N1 −N0 +2)(N1 −N0 )
V ar(X) = 12

3.8
a) Let X be the number that chooses theatre 1, X ∼ binom(n = 1000, p = 0.5).
N is smallest non-negative integer such that
N
1000
1 x
   1000−x
1
P (X ≥ N ) = 1 − −
P
x 2 1 2 < 0.01
i=0

b) X is approximately normally distributed X ∼ N (np, np(1 − p)) = N(500, 250).

N = 536.28 (with continuity correction), i.e. use N = 537 seats.

3.20
q
2
a) E[X] = π , V ar[X] = 1 − π2 .

b) Y = g(X) = X 2 . Y ∼ gamma(α = 1/2, β = 2).

3.23
R∞
a) f (x) ≥ 0 and −∞ f (x)dx = 1.
βα
b) E[X] = β−1 , β > 1.
βα2
E[X 2 ] = β−2 , β > 2.
βα2
V ar[X] = (β−2)(β−1)2
, β > 2.

c Variance doies not exist for β ≤ 2 because E[X 2 ] does not exist.

3.28
2 2)
2 /(2σ 2 ) e−µ√/(2σ
a) σ known: h(x) = e−x , c(µ) = 2π
, t1 (x) = x
σ2
, w1 (µ) = µ.
µ known: h(x) = 1, c(σ) = √ 1 , t1 (x) = −(x − µ)2 , w1 (σ) = 1
.
2πσ 2σ 2

1 1
b) α known: h(x) = xα−1 I{x>0} (x), c(β) = Γ(α)β α , t1 (x) = −x, w1 (β) = β .
β known: h(x) = e−x/β I{x>0} (x), c(α) = Γ(α)β1
α , t1 (x) = ln(x), w1 (α) =
α − 1.
1
α, β unknown: h(x) = I{x>0} (x), c(α, β) = Γ(α)β α , t1 (x) = ln(x), w1 (α, β) =
α − 1, t2 (x) = −x, w2 (α, β) = β1 .
{Ix=0,1,...} (x)
d) h(x) = x! , c(µ) = e−µ , w1 (µ) = ln(µ), t1 (x) = x.

1
3.29
a) σ known: Natural parameter is η1 = µ with natural parameter space {η1 :
−∞ < η1 < ∞}.
µ known: Natural parameter is η1 = σ12 with natural parameter space
{η1 : η1 > 0}.

b) α known: Natural parameter is η1 = β1 with natural parameter space {η1 :

η1 > 0}.
β known: Natural parameter is η1 = α − 1 with natural parameter space
{η1 : η1 > −1}.
α, β unknown: Natural parameter is (η1 , η2 ) = (α − 1, β1 ) with natural
parameter space {(η1 , η2 ) : η1 > −1, η2 > 0}.

d) Natural parameter is η = ln(µ) with natural parameter space is {η : −∞ <

η < ∞}

3.40
X−µ
Can create locations-scale family based on distribution of Z = σ which has
E(Z) = 0, V ar(Z) = 1.
The pdf of Z has pdf fZ (z) = fX (σz + µ) · σ

3.45
R∞ tx R∞
Use that e fX (x)dx ≥ eta fX (x)dx.
∞ a