STATS 100A HW5

Problem 1 Suppose X ∼ f (x), and let Y = aX + b, where a and b are constants.

(1) Prove E(Y ) = aE(X) + b, and Var(Y ) = a2 Var(X).
A:
Z Z Z
E(Y ) = E(aX + b) = (ax + b)f (x)dx = a xf (x)dx + b f (x)dx = aE(X) + b

Var(Y ) = Var(aX + b) = E[aX + b − E(aX + b)]2 = E[aX + b − aE(X) − b]2

= E[a(X − E(X))]2 = a2 E[X − E(X)]2 = a2 Var(X)

(2) Assuming a > 0, calculate the density of Y , g(y).

A: y = ax + b, x = (y − b)/a.
g(y)∆y = f (x)∆x.
∆x/∆y = 1/a.
g(y) = f ((y − b)/a)/a.
Problem 2 Suppose Z ∼ N(0, 1), i.e.,
1 2
f (z) = √ e−z /2 .
2π
(1) Calculate E(Z) and Var(Z).
A:
Z ∞
1 z2 1 z2
E(Z) = z √ e− 2 dz = − √ e− 2 |∞
−∞ = 0.
−∞ 2π 2π

Z ∞
+∞ Z ∞ Z ∞
1 2 1 2 1 −z 2 /2 1 2
√ z 2 e−z /2 dz = − √ ze−z /2  √ e−z /2 dz = 1

Var(Z) = + √ e dz = 0 +
−∞ 2π 2π −∞ −∞ 2π −∞ 2π

(2) Let X = µ + σZ. Calculate the density of X, g(x), E(X), and Var(X) based on Problem 1.
A: Let X = µ + σZ,

1 x−µ 1 (x−µ)2
g(x) = f( )= √ e− 2σ2
σ σ 2πσ 2

i.e., X ∼ N(µ, σ 2 ). Next,

E(X) = 0, Var(X) = σ 2

(3) Suppose P (Z ∈ [−2, 2]) = 95%, then what is P (X ∈ [µ − 2σ, µ + 2σ])?

A:
X −µ
P (X ∈ [µ − 2σ, µ + 2σ]) = P (−2 ≤ ≤ 2) = P (Z ∈ [−2, 2]) = 95%
σ

1
 Problem 3 Poisson process: Suppose we divide the time axis into small periods (0, ∆t),
(∆t, 2∆t), .., (t, t + ∆t), ... Suppose within each interval, we flip a coin independently. Sup-
pose the probability of getting a head is λ∆t. Let T be the time until the first head. Let X be the
number of heads within [0, t].
(1) Find the probability density function of T , and calculate E(T ) based on Geometric distri-
bution.
A:

P (T ∈ (t + ∆t)) (1 − λ∆t)t/∆t (λ∆t)

f (t) = lim = lim = λ lim (1 − λ∆t)t/∆t = λe−λt
∆t→0 ∆t ∆t→0 ∆t ∆t→0

Let T̃ ∼ Geometric(p = λ∆t). T = T̃ ∆t. E(T ) = E(T̃ )∆t = 1/(λ∆t)∆t = 1/λ.

(2) Calculate E(X) based on Binomial distribution.
X ∼ Binomial(n = t/∆t, p = λ∆t). E(X) = np = (t/∆t)(λ∆t) = λt.
Problem 4 Brownian motion or diffusion: Suppose a particle starts from 0, and within each
period, it moves forward or backward by ∆x, each with probability 1/2. Let Xt be the position at
time t (assuming t is a multiple of ∆t). Suppose there are n periods within [0, t], i.e., ∆t = t/n.
Then we can write
Xn
Xt = Zi ∆x,
i=1

where P (Zi = 1) = P (Zi = −1) = 1/2, and Zi are independent.

(1) Calculate E(Xt ) and Var(Xt ).
A:
n
X Xn n
X
E(Xt ) = E( Zi ∆x) = ∆xE( Zi ) = ∆x E(Zi ) = 0
i=1 i=1 i=1
n
X Xn n
X
2 2
Var(Xt ) = Var( Zi ∆x) = (∆x) Var( Zi ) = (∆x) Var(Zi ) = n(∆x)2
i=1 i=1 i=1

(2) Assuming Var(Xt ) = σ 2 t. Then what is the relationship between ∆x and ∆t?
A:
t
Var(Xt ) = n(∆x)2 = σ 2 t =⇒ (∆x)2 = σ 2 = σ 2 ∆t
n
√
Thus ∆x = σ ∆t.
(3) According to the central limit theorem, what is the distribution of Xt ?
A: Xt ∼ N(0, σ 2 t).