Exponential Distribution

Continuous Data Sampling Distribution (STU 07215)

Leguma Bakari
Email: [email protected]
Email: [email protected]
Phone:+255 762 760 095
September 3, 2022
Eastern Africa Statistical Training Center (EASTC)

1
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

2
Definition

• A random variable X that equals the time (or distance, area,

volume) between two successive events (counts) of a Poisson
process with mean 𝜆 > 0 is an exponential random variable with
parameter 𝜆.
• A random variable X is said to follow Exponential distribution if
it satisfies the following density function
•
𝜆e−𝜆x ; x ≥ 0



f ( x; 𝜆) =
 0, otherwise


• Exponential distribution obtains it’s name from the exponential
function in the probability density function .

3
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

4
Moment Generating Function

• Mx ( t ) = E ( etx ) = x etx f ( x ) dx
∫
∫∞ ∫∞ ∫∞
• Mx ( t ) = 0 etx 𝜆e−𝜆x dx = 𝜆 0 etx −𝜆x dx = 𝜆 0 e ( t −𝜆) x dx
h (t −𝜆) x i x =∞ h − (𝜆−t ) x i x =∞
• Mx ( t ) = 𝜆 e t −𝜆 = −𝜆 e 𝜆−t
x =0 x =0
 − (𝜆−t ) x  x =∞
• Mx ( t ) = 𝜆−
−𝜆
t
e x =0
• Mx ( t ) = 𝜆−t e − e = 𝜆−t ( 0 − 1)
−𝜆
 −∞ 0
 −𝜆

• Mx ( t ) = 𝜆
𝜆− t

5
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

6
Expected Value

• E ( X ) = Mx′ ( t )| x =0
• Mx ( t ) = 𝜆
𝜆− t = 𝜆(𝜆 − t ) −1
• Mx′ ( t ) = −𝜆(𝜆 − t ) −2 (−1) = 𝜆(𝜆 − t ) −2
• E ( X ) = Mx′ ( t )| x =0 = 𝜆(𝜆 − 0) −2 = 𝜆𝜆 −2 = 𝜆 −1
• E (X ) = 1
𝜆

7
Variance

• Var ( X ) = E ( X 2 ) − [ E ( X )] 2
• E ( X 2 ) = Mx′′ ( t )| x =0
• Recall; Mx′ ( t ) = 𝜆(𝜆 − t ) −2
• Mx′′ ( t ) = −2𝜆(𝜆 − t ) −3 (−1) = 2𝜆(𝜆 − t ) −3
• E ( X 2 ) = Mx′′ ( t )| x =0 = 2𝜆(𝜆 − 0) −3 = 2𝜆𝜆 −3 = 2𝜆 −2 = 2
𝜆2
 2
• Var ( X ) = E ( X 2 ) − [ E ( X )] 2 = 𝜆22 − 𝜆1 = 𝜆22 − 𝜆12
• Var ( X ) = 1
𝜆2

8
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

9
Distribution Function

∫a
• F ( a) = P ( X < a) = 0 f ( x ) dx
∫a h −𝜆x i x =a
• F ( a) = 0 𝜆e−𝜆x dx = 𝜆 e−𝜆 = [−e−𝜆x ] xx ==a0
x =0
• F ( a) = −[ e −𝜆a − e −𝜆( 0 ) ] = −e −𝜆a + e0
• F ( a) = 1 − e −𝜆a
• Complement of the distribution function
• F ( a) = P ( X > a) = 1 − F ( a)
• F ( a) = 1 − ( 1 − e −𝜆a ) = 1 − 1 + e −𝜆a
• F ( a) = e −𝜆a
• Recall a mass function for Poisson distribution
(𝜆t ) x e −𝜆t
• p (x ) = x!
0 −𝜆t
• At x = 0, then p ( x = 0) = (𝜆t )0!e = e−𝜆t
• Which means there is no count ( x = 0) at e−𝜆t = F ( t )
10
Application of distribution function in computing probabilities

• Application of distribution function in computing probabilities

• P ( X < a) = F ( a) = 1 − e−𝜆a
• P ( a < X < b) = F ( b) − F ( a) = ( 1 − e−𝜆b ) − ( 1 − e−𝜆a )
• P ( a < X < b) = F ( b) − F ( a) = 1 − e−𝜆a − 1 + e−𝜆a = e−𝜆a − e−𝜆b
• Alternatively, using the complement of distribution function
• P ( a < X < b) = F ( a) − F ( b) = e−𝜆a − e−𝜆b

11
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

12
Relationship between Exponential and Poisson Distribution

• An interesting feature of these two distributions is that,

• If the Poisson provides an appropriate description of the number
of occurrences per interval of time, then
• The exponential will provide a description of the length of time
between the occurrences.
• To understand this, consider that, in a Poisson process, if events
occur on average at the rate of 𝜆 per unit of time then
• There will be on average 𝜆t occurrences per t units of time.
• The Poisson distribution describing this process is therefore
(𝜆t ) x e −𝜆t
P (x ) = x!
from which P ( x = 0) = e−𝜆t .
• P ( x = 0) = e−𝜆t is the probability of no occurrences in t units of
time.
13
• Another interpretation of P ( x = 0) = e−𝜆t is that, this is the
probability that the time, T , to the first occurrence is greater that
t , ie P ( T > t ) = P ( x = 0) = e −𝜆t .
• Conversely, the probability that an event does occur during t
units of time is given by P ( T ≤ t ) = 1 − P ( x = 0) = 1 − e−𝜆t .
• Note that, this is a cumulative exponential distribution which,
when differentiated with respect to t , produces the pdf of the
exponential distribution f ( t ) = 𝜆e−𝜆t .

14
Outline

1 Definition

2 Moment Generating Function

3 Expected value and Variance

4 Distribution Function

5 Relationship between Exponential and Poisson Distribution

6 The Memory-Less Property for Exponential Distribution

15
The Memory-Less Property for Exponential Distribution

• Let X be a random variable the follows an Exponential

distribution with parameter 𝜆.
• Suppose we know the probability of X > t where t is the time,
such as P ( X > t ).
• What is the probability that a random variable X is also greater
than some values s + t ? such as P ( X > s + t ).
• That is what we want to know, we want to find
P ( X > s + t | X > t ) =?.
• By using conditional probability theorem we have,
• P ( A | B) = PP( A(∩BB) )
• Now; P ( X > s + t | X > t ) = P { ( X >Ps(+Xt )∩(
>t )
X >t ) }

• but ( X > s + t ) ∩ ( X > t ) = ( X > s + t ), therefore

P ( X >s+t )
• P ( X > s + t | X > t ) = P ( X >t )
• recall the compliment of distribution function
• P ( X > a) = F ( a) = e −𝜆a 16
• Now; P ( X > t ) = F ( t ) = e−𝜆t
• P ( X > s + t ) = F ( s + t ) = e−𝜆( s+t )
• P ( X > s + t | X > t ) = PP( X( X>>s+t )t ) = F ( s+t )
F (t )
e −𝜆( s+t )
• P (X > s + t |X > t ) = e −𝜆t
= e s+t ) − (−𝜆t )
−𝜆(

• P (X > s + t |X > t ) = e −𝜆s −𝜆t +𝜆t

=e −𝜆s

• P (X > s + t |X > t ) = e −𝜆s

= F ( s) = P ( X > s)
• P ( X > s + t | X > t ) = P ( X > s)
• It turns out that the conditional probability does not depend on t .
• The probability of an exponential random variable exceeding the
value s + t given t is the same as the variable originally exceeding
that value s regardless of t .
• The Exponential distribution is memory-less because the past has
no bearing on the future behaviors.
• Every instant is like beginning a new random period which has
the same distribution regardless of how much time has already
elapsed.
• The Exponential distribution is the only memory-less continuous 17
Practical Examples

Recall EXERCISES FOR SECTION 4-9, page 127 download this

book from here.

18