Scribd
Upload
25 views4 pages

Stok 2 Januar 2019

This document summarizes a reexam for a stochastic processes course given on January 31, 2019. It consists of 4 problems and 13 questions testing concepts related to convergence of random variables and stochastic processes. Students had 4 hours to complete the exam and were allowed all aids except laptops. The questions covered topics like almost sure and L2 convergence of sums and products of independent random variables, asymptotic normality, and the central limit theorem.

Uploaded by

helwegovesen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
25 views4 pages

Stok 2 Januar 2019

This document summarizes a reexam for a stochastic processes course given on January 31, 2019. It consists of 4 problems and 13 questions testing concepts related to convergence of random variables and stochastic processes. Students had 4 hours to complete the exam and were allowed all aids except laptops. The questions covered topics like almost sure and L2 convergence of sums and products of independent random variables, asymptotic normality, and the central limit theorem.

Uploaded by

helwegovesen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

UNIVERSITY OF COPENHAGEN

Stochastic Processes 2 (Stok2)


Reexam 2018-19
January 31th 2019

4 hours ITX exam. All aids are allowed, except for laptops and similar devices.

The problem set consists of 4 pages. There are 3 problems with a total of 13
questions. The 13 questions carry the same weight in the final grading.

Note the convention that all limits are understood as for n → ∞ if nothing
else is stated.

Problem 1

Let X1 , X2 , . . . be independent real-valued random variable. We √


assume√for
each n ∈ N that Xn is concentrated on the three-point set {− n, 0, n}
with proint probabilites
√
P (Xn = − n) = n12
P (Xn = 0) = 1 − n22
√
P (Xn = n) = n12

Question 1.1. Show that E Xn2 < ∞ for all n, and compute
E Xn and E Xn2 .

Pn
Consider Sn = i=1 Xi .

Question 1.2. Show that there is a random variable S such that


Sn → S almost surely.

Question 1.3. Will Sn → S in L2 ?

1
Question 1.4. Will Sn → S in L1 ?

√
Question 1.5. Show that Sn / n → 0 almost surely and in L2 .

Pn
Consider Tn = i=1 Xi Xi+1 .

Question 1.6. Show that there is a limit variable T such that


Tn → T almost surely and in L2 .
Warning: the terms in Tn may not be independent.

Problem 2

Let X1 , X2 , . . . , Y1 , Y2, . . . be independent real-valued random variables, all


having a standard exponential distribution. Consider
v ! n !
u n
1u X X
Tn = t Xi Xi Y i
n i=1 i=1

Question 2.1. Show that Tn → 1 almost surely.

Question 2.2. Find an asymptotic normal distribution for


n  
1X Xi
n i=1 Xi Yi

Question 2.3. Show that

β2
 
a.s.
Tn ∼ N α,
n

and identify the asymptotic parameters α and β 2 .

2
Problem 3

Let X1 , X2 , . . . be independent. They are not necessarily identically distribu-


ted, but we assume that E Xn = 0 for all n ∈ N and that there is a constant
C > 0 such that
P (|Xn | ≤ C) = 1 for all n ∈ N
Consider n
X Sn
Sn = Xi and Yn = √
i=1
V Sn

Question 3.1. Assume for this question that there are constants
α > 0 and K > 0 such that

V Sn ≥ Knα for all n ∈ N


D
Use Lyapounovs CLT to show that Yn → N (0, 1). Hint: What
order of the Lyapounov condition is relevant for the given value
of α?

Question 3.2. Assume for this question that


∞
X
E Xn2 = ∞ (1)
n=1

D
Use Lindebergs CLT to show that Yn → N (0, 1).

Question 3.3. Assume for this question that


∞
X
E Xn2 < ∞ (2)
n=1

D
Show that there is a limit variable Y such that Yn → Y .

Question 3.4. Assume for this question that


   
1 1 1
P Xn = n = P Xn = − n =
3 3 2

3
Show that (2) is satisfied. Hence it is known from the previous
D
question that there is a limit variable Y such that Yn → Y . Show
that √ !
2
P |Y | < =0
3
and conclude that Y is not N (0, 1)-distributed.

You might also like