5 Questions in Final Exam - Theory of Probability | MATH 831, Exams of Probability and Statistics

Material Type: Exam; Professor: Valko; Class: Theory of Probability; Subject: MATHEMATICS; University: University of Wisconsin - Madison; Term: Fall 2008;

Typology: Exams

2010/2011

Uploaded on 07/25/2011

koofers-user-pvx
koofers-user-pvx 🇺🇸

10 documents

1 / 1

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Math/Stat 831 Fall 2008
Final Exam
1. (9 points) Let X1, X2, . . . be i.i.d. with mean 0 and variance 1. Let anbe an increasing sequence of
positive numbers, and denote An= (Pn
k=1 a2
n)1/2. Show that if an/An0 then
n
P
k=1
akXk
An
N(0,1),
i.e. the normalized sums converge to a standard normal random variable in distribution.
2. (10 points) Suppose that for two random variables Xand Ythe following statements hold:
XYand Yare independent,
XYand Xare independent.
(a) Show that if the characteristic functions of Xand XYare denoted by ψX(t) and ψXY(t) then
ψX(t) (1 |ψXY(t)|2) = 0.(1)
(b) Show that if equation (1) holds for all tthen |ψXY(t)|= 1 if tis close enough to 0.
(c) Show that XYis constant with probability 1.
3. (10 points) Let ξnbe independent random variables with distribution
ξn=
1 with probability (2n)1
0 with probability 1 1/n
1 with probability (2n)1
Let X0= 0 and define Xnrecursively by
Xn+1 =|ξn|nXn+ξn1(Xn= 0).
(a) Show that Xnis a martingale.
(b) Show that Xn
P
0. (Hint: it’s easy to estimate P(Xn= 0). . . )
(c) Show that Xndoes not converge to 0 a.s. (Hint: then ξnshould be 0 eventually.)
4. (9 points) Consider the Lebesgue probability space on the interval [0,1). (I.e. the state space is
= [0,1), the σ-field is the set of Lebesgue measurable sets and the measure is the Lebesgue measure.)
We define the random variable Xas
X(ω) = (2ω, if 0 ω < 1/2,
2ω1 if 1/2ω < 1.
Compute the conditional expectation E(Y|X) where Y: [0,1) Ris a measurable function.
5. (12 points) Let X1, X2, . . . be i.i.d. with P(Xn= 1) = p, P(Xn=1) = 1 pwhere 1/2< p < 1 is
a fixed constant. Let Sn=Pn
k=1 Xn(with S0= 0) and ϕ(x) = (1/p 1)x.
(a) Show that ϕ(Sn) is a martingale with respect to the natural filtration.
(b) For given a, b positive integers consider the first time Snreaches aor b:
τ= inf{k:Sk=aor Sk=b}.
Show that τis a stopping time with respect to the natural filtration.
(c) Show that Eϕ(Sτ) = 1 and compute the probability of reaching abefore b, i.e. P(Sτ=a).
(d) Find a suitable martingale to compute Eτ. (Hint: what is the expectation of Xn?)
Please present your solutions in a clear manner. Show all your work.

Partial preview of the text

Download 5 Questions in Final Exam - Theory of Probability | MATH 831 and more Exams Probability and Statistics in PDF only on Docsity!

Math/Stat 831 – Fall 2008

Final Exam

  1. (9 points) Let X 1 , X 2 ,... be i.i.d. with mean 0 and variance 1. Let an be an increasing sequence of positive numbers, and denote An = (

∑n k=1 a

2 n) 1 / (^2). Show that if an/An → 0 then

∑^ n k=

akXk

An

⇒ N (0, 1),

i.e. the normalized sums converge to a standard normal random variable in distribution.

  1. (10 points) Suppose that for two random variables X and Y the following statements hold:
    • X − Y and Y are independent,
    • X − Y and X are independent.

(a) Show that if the characteristic functions of X and X − Y are denoted by ψX (t) and ψX−Y (t) then ψX (t) (1 − |ψX−Y (t)|^2 ) = 0. (1)

(b) Show that if equation (1) holds for all t then |ψX−Y (t)| = 1 if t is close enough to 0. (c) Show that X − Y is constant with probability 1.

  1. (10 points) Let ξn be independent random variables with distribution

ξn =

1 with probability (2n)−^1 0 with probability 1 − 1 /n − 1 with probability (2n)−^1 Let X 0 = 0 and define Xn recursively by Xn+1 = |ξn|nXn + ξn 1 (Xn = 0). (a) Show that Xn is a martingale. (b) Show that Xn P −→ 0. (Hint: it’s easy to estimate P(Xn = 0)... ) (c) Show that Xn does not converge to 0 a.s. (Hint: then ξn should be 0 eventually.)

  1. (9 points) Consider the Lebesgue probability space on the interval [0, 1). (I.e. the state space is Ω = [0, 1), the σ-field is the set of Lebesgue measurable sets and the measure is the Lebesgue measure.) We define the random variable X as

X(ω) =

2 ω, if 0 ≤ ω < 1 / 2 , 2 ω − 1 if 1/ 2 ≤ ω < 1.

Compute the conditional expectation E(Y |X) where Y : [0, 1) → R is a measurable function.

  1. (12 points) Let X 1 , X 2 ,... be i.i.d. with P(Xn = 1) = p, P(Xn = −1) = 1 − p where 1/ 2 < p < 1 is a fixed constant. Let Sn =

∑n k=1 Xn^ (with^ S^0 = 0) and^ ϕ(x) = (1/p^ −^ 1) x.

(a) Show that ϕ(Sn) is a martingale with respect to the natural filtration. (b) For given a, b positive integers consider the first time Sn reaches a or −b: τ = inf{k : Sk = a or Sk = −b}. Show that τ is a stopping time with respect to the natural filtration. (c) Show that Eϕ(Sτ ) = 1 and compute the probability of reaching a before b, i.e. P(Sτ = a). (d) Find a suitable martingale to compute Eτ. (Hint: what is the expectation of Xn?)

Please present your solutions in a clear manner. Show all your work.