STAT8102 PROBABILITY

ASSIGNMENT

Question 1:

Suppose that X and Y are continuous random variables with joint pdf given
by
(
c x2 y + 2y 2 , 0 < x < 2, 0 < y < 1,


f (x, y) =
0, otherwise.
(a) Find the value for c that makes f (x, y) a valid pdf.
(b) Determine the joint odf of X and Y on the  entire xy-plane.
(c) Use the cdf to determine P {1 < X < 2} ∩ 12 < Y < 1 .

(d) Determine the marginal pdf fX (x) of X and the marginal pdf fY (y) of Y .
(e) Determine the conditional pdf gX (x | y).
(f ) Determine the probability that X > 1 given Y = 12 .

Solution:

(a) To find the value of c that makes the given function a valid joint proba-
bility density function (pdf ):

For this we need to ensure that the total probability over the entire range is equal to 1.
Mathematically, we have:

Z 2Z 1
c(x2 y + 2y 2 ) dy dx = 1
0 0
First, integrate with respect to y from 0 to 1:

2 1
x2 2 2 3
Z 
c y + y dx = 1
0 2 3 0
Now, evaluate the inner part:

c 2 2c
x + =1
2 3
1
2 ASSIGNMENT

To find c, we need to satisfy this equation for all x in the range [0, 2]. We can do this
by integrating with respect to x from 0 to 2:

Z 2 
c 2 2c
x + dx = 1
0 2 3
c 3 2c 2
x + x =1
6 3 0

8c 4c
+ =1
3 3
12c
=1
3
4c = 1

1
c=
4
So, the value of c that makes f (x, y) a valid pdf is c = 14 .

(b) The joint pdf of X and Y on the entire xy-plane is:

(
1 2
4 (x y + 2y 2 ), 0 < x < 2, 0 < y < 1
f (x, y) =
0, otherwise
(c) To find P ({1 < X < 2} ∩ { 21 < Y < 1}):

Bt using the cumulative distribution function (CDF). First, find the CDF of the joint
distribution:

Z y Z x
1 2
F (x, y) = (u v + 2v 2 ) du dv
0 0 4
Now, calculate the probability:

1
P ({1 < X < 2} ∩ { < Y < 1}) = F (2, 1) − F (1, 1) − [F (2, 1/2) − F (1, 1/2)]
2
You can substitute the CDF values into this equation to find the probability.

(d) To determine the marginal pdf of X, integrate the joint pdf over the
entire range of y:
 STAT8102 PROBABILITY 3

Z 1
1 2
fX (x) = (x y + 2y 2 ) dy
0 4

1 x2 2 2 3 1
 
fX (x) = y + y
4 2 3 0
 2 
1 x 2
fX (x) = +
4 2 3
Simplify:

x2 1
fX (x) = + , for 0 < x < 2
8 6
To determine the marginal pdf of Y , integrate the joint pdf over the entire range of x:

Z 2
1 2
fY (y) = (x y + 2y 2 ) dx
0 4
 
1 1 3 2 2 2
fY (y) = xy + y
4 3 3 0
 
1 2 3 2 2
fY (y) = y + y
4 3 3
Simplify:

1
fY (y) = y 2 (2y + 1), for 0 < y < 1
6
(e) To determine the conditional pdf gX (x|y), you can use the conditional
probability formula:

f (x, y)
gX (x|y) =
fY (y)
Substitute the expressions for f (x, y) and fY (y):

1 2 2
4 (x y + 2y )
gX (x|y) = 1 2
6 y (2y + 1)
Simplify:

3x2
gX (x|y) =
8(2y + 1)
4 ASSIGNMENT

(f ) To find the probability that X > 1 given Y = 21 :

By using the conditional probability formula:

Z 2
1 1
P (X > 1|Y = ) = gX (x| ) dx
2 1 2
Substitute the expression for gX (x|y) when y = 12 :

2
3x2
Z
1
P (X > 1|Y = ) = dx
2 1 8(2( 12 ) + 1)
Simplify:

2
3x2
Z
1
P (X > 1|Y = ) = dx
2 1 8
Now, calculate the integral:

x3
 
1 3 2
P (X > 1|Y = ) =
2 8 3 1
 
1 3 8 1
P (X > 1|Y = ) = −
2 8 3 3
1 7
P (X > 1|Y = ) =
2 8
Question 2:

Suppose that X and Y have a continuous joint distribution with joint pdf
(
c x2 + y , 0 ≤ y ≤ 1 − |x|


f (x, y) =
0, otherwise.
Determine the correlation ρ(X, Y ).

Solution:

To determine the correlation, ρ(X, Y ), we first need to find the covariance and the stan-
dard deviations of the random variables X and Y.

The covariance, denoted as Cov(X, Y ), can be calculated as follows:

 STAT8102 PROBABILITY 5

Cov(X, Y ) = E[XY ] − E[X]E[Y ]

Where:
- E[XY ] is the expected value of the product of X and Y.
- E[X] and E[Y ] are the expected values (means) of X and Y, respectively.

To calculate E[XY ], we’ll use the joint probability density function (pdf):

Z ∞ Z ∞
E[XY ] = xyf (x, y) dx dy
−∞ −∞
In your case, the limits of integration are determined by the support of the joint pdf,
which is defined as:

0 ≤ y ≤ 1 − |x|
So, the integral becomes:

Z 1 Z 1−|x|
E[XY ] = xyc(x2 + y) dy dx
−1 0
Now, we need to find E[X] and E[Y ], which are the expected values of X and Y:

Z ∞
E[X] = xfX (x) dx
−∞
Z ∞
E[Y ] = yfY (y) dy
−∞
To find the marginal pdfs, we integrate the joint pdf over the other variable:

For X: Z 1−|x|
fX (x) = c(x2 + y) dy
0
For Y: Z 1
fY (y) = c(x2 + y) dx
−1
After finding these marginal pdfs, you can calculate E[X] and E[Y ] by taking the ex-
pected values using the marginal pdfs.

Once you have computed E[XY ], E[X], and E[Y , you can use the formula for the cor-
relation coefficient ρ(X, Y ):
6 ASSIGNMENT

Cov(X, Y )
ρ(X, Y ) =
σX σY
Where σX and σY are the standard deviations of X and Y, respectively. The standard
deviations can be calculated as:

p p
σX = Var(X) = E[X 2 ] − (E[X])2

p p
σY = Var(Y ) = E[Y 2 ] − (E[Y ])2
So, the correlation coefficient can be computed once you have these values.

Question 3:

Suppose you toss a fair coin independently 30 times. Use the Central Limit
Theorem to approximate the probability that the number of heads is one of
the following.
(a) Greater than 20 .
(b) Greater than or equal to 10 and less than 15.
(c) How do the values found compare to the exact values? Do you expect these
to be the same?

Solution:

To approximate the probabilities using the Central Limit Theorem (CLT), we will use
the properties of the binomial distribution. When you toss a fair coin independently 30
times, you have a binomial distribution with parameters n (number of trials) and p (prob-
ability of success, which is the probability of getting heads). In this case, n = 30, and p =
0.5, as you mentioned that the coin is fair.

(a) Probability of getting greater than 20 heads:

Using the CLT, we can approximate this probability by treating the binomial distribu-
tion as a normal distribution with the same mean and variance. The mean is given by
np, and the variance is given by npq, where q is the probability of failure (probability of
getting tails).

µ = np = 30 ∗ 0.5 = 15σ 2 = npq = 30 ∗ 0.5 ∗ 0.5 = 7.5

p √
Now, we can calculate the standard deviation as σ = (σ 2 ) = 7.5≈ 2.74.
 STAT8102 PROBABILITY 7

Now, we want to find the probability of getting more than 20 heads, which corresponds
to X > 20. We standardize this value using the z-score formula:

(20 − 15)
Z = f rac(X − µ)σZ = ≈ 1.82
2.74
Now, we find the probability using the standard normal distribution table or calculator:

P (Z > 1.82) ≈ 0.0344

So, the probability of getting greater than 20 heads is approximately 0.0344.

(b) Probability of getting greater than or equal to 10 and less than 15 heads:

Using the CLT, we can approximate these probabilities as follows:

1. For greater than or equal to 10 heads:

(10 − 15)
Z1 = ≈ −1.82P (Z > −1.82) ≈ 0.9656
2.74
2. For less than 15 heads:
(15 − 15)
Z2 = ≈ 0.00P (Z < 0.00) ≈ 0.5000
2.74
Now, to find the probability of getting greater than or equal to 10 and less than 15
heads, we subtract:

P (10 ≤ X < 15) = P (Z2) − P (Z1) ≈ 0.5000 − 0.9656 ≈ −0.4656

(c) Comparison to Exact Values:

The Central Limit Theorem provides approximations, and these values should be rea-
sonably close to the exact values, especially when the number of trials (n) is sufficiently
large. The exact values can be calculated using the binomial distribution directly, but the
CLT provides a quick and often accurate approximation.

The CLT is more accurate when n is large because it relies on the normal distribution,
which becomes a better approximation as the sample size increases. In this case, with n =
30, the CLT approximations should be quite close to the exact values, but there may still
be some small differences due to the discrete nature of the binomial distribution. Overall,
for practical purposes, the CLT approximations are often considered accurate enough,
especially when n is sufficiently large.