X=0 X=1 X=2
Y=0 0,17 0,11 0,05 0,17+0,11+0,05=
0,33
Y=1 0,08 0,45 0,14 0,08+0,45+0,14=
0,67
0,17+0,08= 0,25 0,56 0,19 1
E ( X )=0 , 25· 0+ 0 ,56 · 1+0 , 19· 2=0 ,94
E ( X )=0 , 94
E ( Y )=0· 0 , 33+0 , 67 · 1=0 , 67
2
V ( X ) ¿ ( x 1−E ( X ) ) · f ( x 1 )
2 2 2
V ( X )= ( 0−0 , 94 ) · 0 , 25+ ( 1−0 , 94 ) ·0 , 56+ ( 2−0 , 94 ) · 0 ,19=0,4364
V ( X )=0,4364
Other way:
V ( X )=E ( X 2 )−E ( X )
2
2 2 2 2
E( X )=0 · 0.25+1 ∗0.56+2 ∗0.19=1 , 32
2 2
E ( X ) =0.94 =0,8836
2
V ( X )=1.32−0.94 =0,4364
2
V ( Y ) ¿ ( y 1−E ( Y ) ) · f ( y 1 )
V ( Y )=¿
V ( Y )=0,2211
Std ( X )= √ 0,4364 ≈ 0,660605783
Std ( X ) ≈ 0,660605783
Std ( Y )=√ 0,2211≈ 0,470212718
Std ( Y ) ≈ 0,470212718
Cov ( X , Y ) =E ( XY )−E ( X ) · E(Y )
E ( XY ) =1· 1· 0 , 45+ 2· 1· 0 , 14=0 , 73
E ( X ) · E (Y )=0 , 94 · 0 , 67=0,6298
Cov ( X , Y ) =0 ,73−0,6298=0,1002
Cov ( X , Y ) =0,1002
Cov( X ,Y )
Corr ( X ,Y )=
√V (X )· √ V ( Y )
0,1002
Corr ( X ,Y )= ≈ 0,322575203
√ 0,4364 · √0,2211
Corr ( X ,Y ) ≈ 0,322575203
0 , 56 · 0 ,67 ≈ 0,3752 ≠ 0 , 45
Thus, they are dependent.
X=0 X=1 X=2
Y=0 0 1 2
Y=1 1 2 3
Z=0 Z=1 Z=2 Z=3 Sum
0,17 0 , 08+0 , 11=0 , 190 , 45+ 0 , 05=0 , 5 0,14 0,14+0,5+0,19+0,17= 1
E ( Z )=0 · 0 ,17+1 · 0 , 19+ 2· 0 , 5+3 · 0 ,14=1 , 61
2 2 2
V ( Z ) ¿ ( 0−1 , 61 ) · 0 , 17+ ( 1−1, 61 ) ·0 , 19+ ( 2−1, 61 ) ·0 , 5+¿
Std ( Z )= √ 0,8579 ≈ 0,926228913
Exercise 2 in the R file
Ex. 2 g)
Exercise 3 a, b, d, e in the R file
Exercise 4 in the R file
Exercise 3 c)
W =10 X +5 Y +Z
Since X, Y , Z are independent and normally distributed, then we know that their sum, i.e. W is also
normally distributed. To know the distribution of W, we only need to know its mean and variance.
E ( W ) =E ( 10 X +5 Y +Z )=10 E ( X )+ 5 E (Y )+ E ( Z )=10 · 0 , 05+5 ·0 , 15+0 , 50=1, 75
2 2
V ( W )=V ( 10 X +5 Y + Z )=10 ·V ( X )+5 · V (Y )+ V ( Z )=¿
¿ 100 · 0,025+25 ·0 , 1+1=6
Std ( W )= √ 6 ≈ 2,44948974
W ∼ N (1.75 , 6)
Exercise 4 b,c,d
b)
For a p ∈ [0, 1], the p-quantile q p(X) is the number such that P (X ≤ q p (X)) = p. For instance, the
probability that X has an outcome less than 1.2816 is 90%. Similarly, there is a 1% probability of an
outcome of Y being less than or equal to 65.1048.
c)
Since the normal distribution is symmetric around its mean, then we have μ X − q p (X) = q 1− p (X ) − μ X
and similarly for Y , μY − q p (X) = q 1− p (Y ) − μY . This means that qp(·) is as far from μ· as q1−p(·) is
from μY . In particular, for the important case of the standard normal distribution X, this means that:
−qp(X) = q1−p(X).
d)
This is straight forward to verify. For instance, for p = 0.01 = 1%, we have
μY +σY · q0 , 01 ¿
which is equal to q 0 ,01 (Y )
