Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Math 7360-2
Homework
1. State the probability mass functions of the distributions: (a) binomial, (b) Poisson,
and (c) geometric.
2. Derive the means and variances of the distributions in Problem 1.
3. State the density functions of the distributions: (a) normal, (b) exponential, and (c)
gamma.
4. Derive the means and variances of the distributions in Problem 3.
5. Let π be a standard normal random variable. Deο¬ne a random variable π by
{
π, if π(π) β₯ 0,
π (π) =
π, if π(π) < 0,
where π, π are two constants. Let π = π + π .
(1) Find the density function of π.
(2) Use part (1) to ο¬nd the mean and variance of π.
(3) Find the covariance of π and π .
6. Let π be a random variable with distribution function given by the Cantor function.
Find the mean and variance of π.
7. Let π (π‘) be a continuous function on [π, π]. Prove that the function
β«
π₯
πΉ (π₯) =
π (π‘) ππ‘,
π β€ π₯ β€ π,
π
is absolutely continuous.
(Note: The conclusion is true for an integrable function π (π‘). But the proof is harder.)
8. Suppose the joint density function of π and π is given by
[
(
)]
1
1
(π₯βπ1 )2
π₯βπ1 π¦βπ2 (π¦βπ2 )2
β
π (π₯, π¦) =
exp β
β 2π
+
,
2(1βπ2 )
π12
π1
π2
π22
2ππ1 π2 1βπ2
where π1 , π2 are real numbers, π1 , π2 are positive numbers, and β1 < π < 1.
(1) Show that π and π have distributions π (π1 , π12 ) and π (π2 , π22 ), respectively,
(2) Show that π is the correlation coeο¬cient of π and π .
(3) Find the density function of π + π .
9. Prove the following equality for any 0 < π < π‘ < π’,
[
]
β« β
1
1 ( π₯2
(π¦ β π₯)2
(π§ β π¦)2 )
β
exp β
+
+
ππ¦
2 π
π‘βπ
π’βπ‘
(2π)3 π (π‘ β π )(π’ β π‘)
ββ
[
]
1 ( π₯2
(π§ β π₯)2 )
1
exp β
+
.
=β
2 π
π’βπ
(2π)2 π (π’ β π )
1
10. Let β³ be the space of random variables on a probability space. Show that
π1 (π, π ) = πΈ
β£π β π β£
1 + β£π β π β£
and π2 (π, π ) = πΈ(β£π β π β£ β§ 1)
deο¬ne two metrics on β³.
11. Let π1 and π2 be the metrics in Problem 17. Check whether there exist constants
πΌ, π½ > 0 such that πΌπ2 (π, π ) β€ π1 (π, π ) β€ π½π2 (π, π ) for all π, π β β³.
12. Prove the equivalence: (1) ππ β π in prob., (2) π1 (ππ π) β 0, (3) π2 (ππ π) β 0.
13. Let {πΈπ } be a sequence of events and let ππ = π (πΈπ ). Find the necessary and
suο¬cient condition on the sequence {ππ } such that 1πΈπ β 0 in probability.
14. Let {πΈπ } be a sequence of independent events and let ππ = π (πΈπ ). Find the necessary
and suο¬cient condition on the sequence {ππ } such that 1πΈπ β 0 almost surely.
15. Let π be a standard normal random variable and let ππ be the truncation of π at
level π > 0. Find the mean and the variance of ππ .
16. Let {ππ } be a sequence of independent random variables with the same uniform
distribution on the interval [0, 1]. Suppose π (π₯) is a continuous function on [0, 1].
Investigate the limit
}
1{
π (π1 ) + π (π2 ) + β
β
β
+ π (ππ )
πββ π
lim
in πΏ1 -convergence, almost sure convergence, and convergence in probability.
17. Let {ππ } be a sequence of independent and identically distributed random variables.
Assume that ππ are nonconstant. Prove that π {π; ππ (π) converges} = 0.
18. Prove the following equality
β«
β
0
( sin π₯ )2
π₯
ππ₯ =
π
.
2
19. Let π (π₯) = sinπ₯ π₯ , π₯ > 0. Prove that π ββ πΏ1 (0, β). However, prove that the following
improper integral exists and has the value
β« β
sin π₯
π
ππ₯ = .
π₯
2
0
20. Let {ππ } be a sequence of independent random variables with the distributions
π (ππ = π) = π (ππ = βπ) =βππ , π (ππ = 0) = 1 β 2ππ , 0 < ππ < 21 , π β₯ 1.
Find conditions on {ππ } so that π ππ converges almost surely.
21. Suppose {ππ } is a sequence of independent random variables with the same distribution
π
β(ππ 1 = 1) = π (ππ = β1) = 1/2. Find the condition on the constant πΌ so that
π ππΌ ππ converges almost surely.
2
22. Let {ππ } be a sequence of independent random variables having the same exponential
β
distribution with parameter π > 0. Find conditions on constants ππ so that π ππ ππ
converges almost surely.
β
β
23. Prove or disprove the statement: If π πΈβ£ππ β£ < β, then π ππ converges absolutely
almost surely.
24. Let π be uniformly distributed on the interval [β1, 1]. Show that the characteristic
sin π‘
function of π is given by π(π‘) =
.
π‘
25. Let ππ be the Gaussian measure with mean ππ and variance ππ2 . Find conditions on
ππ and ππ such that the family {ππ } is tight.
26. Let {ππ } be independent Poisson random variables, each with parameter 1. By
applying the central limit theorem to this sequence, prove that
π
1 β ππ
1
lim
= .
πββ ππ
π!
2
π=0
27. Let {ππ }β
π=1 be a sequence of independent random variables with the distributions
π1 βΌ π (0, 1) and ππ βΌ π (0, 2πβ2 ), π β₯ 2. Let
πππ = ββπ
ππ
π=1
Var(ππ )
,
1 β€ π β€ π.
Show that the triangular array {πππ } does not satisfy the Lindeberg condition.
28. Check whether the binomial distribution π(1, π) is stable.
29. Let ππ be binomial with parameter (π, ππ ) and suppose πππ β π > 0. Prove that
ππ converges in distribution to the Poisson distribution with parameter π.
30. Find the LeΜvy components of a compound Poisson distribution. (The characteristic
function of such a distribution is given by Ξ¦(π‘) = ππ(π(π‘)β1) , π > 0, π(π‘) = πΈπππ‘π1 .)
31. Find the LeΜvy components of a symmetric stable distribution. (The characteristic
π
function of such a distribution is given by π(π‘) = πβπβ£π‘β£ , π > 0, 0 < π β€ 2.)
32. State ten important theorems in this course. For each theorem, give examples and
counterexamples.
3