RWTH Aachen University
Institut für Statistik
und Wirtschaftsmathematik
Prof. Dr. E. Cramer, M. Döring,
T. v. Bentum
https://www.isw.rwth-aachen.de/
Dec. 10th, 2020
Exercise Sheet 7
Mathematics of Data Science
Presented in Tutorial: Dec. 17th, 2020
Lectures: Dec. 10th & Dec. 15th 2020
Topics: Bivariate discrete distributions, conditional probabilities of discrete random variables, convolution
of discrete random variables, independence of random variables, cumulative distribution function
of a random vector, binomial distribution, Poisson distribution, multinomial distribution.
Exercise P 6
Let (X, Y )0 be a bivariate discrete random vector on a probability space (Ω, A, P ) with P (X ∈ {−1, 0, 1}) =
P (Y ∈ {1, 2, 3}) = 1. The corresponding probabilities P (X = i , Y = j) for i ∈ {−1, 0, 1} , j ∈ {1, 2, 3} are
given by the following contingency table:
j
P (X = i , Y = j)
1 2 3
1 1
-1 0
20 5
1 1 1
i 0
5 5 10
1 1 1
1
10 10 20
(a) Determine the corresponding marginal probability mass functions f X of X and f Y of Y . Decide,
whether the random variables X and Y are stochastically independent or not.
(b) Calculate the conditional probability P (X = 0 | Y ≥ 2) .
(c) Determine the distribution of the random variable Z = X + Y .
Exercise P 7
Let X and Y be two stochastically independent random variables on a probability space (Ω, A, P ) with
X ∼ bin(5, 21 ) and Y ∼ po(2) (i.e., X has a binomial distribution with parameters n = 5 and p = 12 and Y
has a Poisson distribution with parameter λ = 2).
(a) Determine the (joint) distribution P (X,Y ) of the bivariate (discrete) random vector (X, Y )0 and the
support of the probability mass function P (X,Y ) .
(b) Let F (X,Y ) denote the (joint) cumulative distribution function of (X, Y )0 according to Definition P
1.50.
Calculate F (X,Y ) (1, 1) .
1/2
Exercise P 8
Prove the statement of Theorem P 2.36, i.e.:
Let λ1 , . . . , λk > 0 and X1 , . . . , Xk be stochastically independent random variables on a probability space
(Ω, A, P ) with Xi ∼ po(λi ) for i ∈ {1, . . . , k}, i.e. Xi has a Poisson distribution with parameter λi , i ∈
{1, . . . , k}.
Further, let X := (X1 , . . . , Xk )0 and
λi
pi := Pk for i ∈ {1, . . . , k} .
j=1 λj
Then, for n ∈ N, it holds: Pk
PX | i=1 Xi = n
= M (n, p1 , . . . , pk ) .
(Given that the sum of the components of X = (X1 , . . . , Xk )0 is equal to n ∈ N, X has a multinomial
distribution with parameters n and p1 , . . . , pk as given above.)
Exercise P 9
In a shipment of 200 Smart TVs there are 4 TVs with some irreparable defect (i.e. these units have to be
returned to the manufacturer), 16 TVs with some (slight) defect that could be repaired by the corresponding
retailer, and the remaining TVs have no defect.
For quality control, the recipient of the shipment proceeds as follows:
One after another, 10 TVs are selected from the shipment with replacement and checked for defects (after
which the selected unit is returned to the shipment).
(a) Calculate the probability, that the sample of 10 TVs contains exactly 1 TV with some failure that
could be repaired by the retailer and 9 TVs without any defect.
(b) Calculate the probability that the sample of 10 TVs contains at least 8 TVs without any defect.
2/2
