Scribd
Upload
44 views

Lecture15 PDF

The document discusses the gamma and exponential distributions. It provides the probability density functions, cumulative distribution functions, moment generating functions, and formulas for the mean and variance of both distributions. It also proves that the waiting time between events in a Poisson process has an exponential distribution.

Uploaded by

Apoorav Dhingra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
44 views

Lecture15 PDF

The document discusses the gamma and exponential distributions. It provides the probability density functions, cumulative distribution functions, moment generating functions, and formulas for the mean and variance of both distributions. It also proves that the waiting time between events in a Poisson process has an exponential distribution.

Uploaded by

Apoorav Dhingra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 29

MATH F113

(Probability and Statistics)


Chandra Shekhar
Associate Professor

Department of Mathematics
BITS Pilani, Pilani Campus, Rajasthan 333 031
Email: [email protected]
Mobile: 9414492349
Chandra Shekhar MATH F113 (Probability and Statistics)
What have you covered?

In Lecture 14
Exercise Problem
Gamma Function

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution

Gamma Distribution A random


variable X with density function
( −x
1 α−1 β
αx e x > 0 α > 0, β > 0
f (x) = Γ(α)β
0 elsewhere
is said to have a Gamma Distribu-
tion with parameters α, β for α > 0, β >
0

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

To check the necessary and sufficient


condition of pdf: f (x) ≥ 0 for all x > 0
Further
Z∞ Z∞
1 x
α−1 − β
f (x)dx = x e dx
Γ(α)β α
−∞ 0

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Let
x
=t
β
dx = βdt
and
x = βt
Z∞
1
= α
β α−1tα−1e−tβdt
Γ(α)β
0

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

α Z∞
β
tα−1e−tdt = 1
Γ(α)β α
0
Hence f (x) is a p.d.f

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Theorem: Let X be a gamma


random variable with parame-
ter α & β, then m.g.f for X is
given by
mx(t) = (1 − βt)−α
Hence,
E [X] = αβ
V ar(x) = αβ 2
Chandra Shekhar MATH F113 (Probability and Statistics)
Gamma Distribution (Cont...)

Z ∞
tx
mx(t) = E[e ] = etxf (x)dx
Z ∞ −∞
1
= etx α e−x/β xα−1dx
0 β Γ(α)
Z ∞
1
= α e−(1/β−t)xxα−1dx
β Γ(α) 0

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Let z = ( β1 −t)x, or z = (1−βt) βx ⇒ x =


zβ βdz
(1−βt) and dx = (1−βt)
Z ∞ α−1 −z
β e z α−1 βdz
mx(t) =
0 β α Γ(α) (1 − βt)α−1 (1 − βt)
Z ∞
1 −z α−1
= e z dz
(1 − βt)α Γ(α) 0
1
= Γ(α)
(1 − βt)α Γ(α)

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Since,
Z ∞
e−z z α−1dz = Γ(α)
0

Therefore,
1 1
mx(t) = , t< .
(1 − βt)α β

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

 
d
E[X] = mx(t)
dt t=0
= −α(1 − βt)−α−1(−β) t=0 = αβ
 

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

2
 
d
E[X 2] = (mx(t))
dt2 t=0
 
d
= (−α(1 − βt)−α−1(−β))
dt t=0
= αβ(−α − 1)(−β)

V ar(X) = αβ(−α − 1)(−β) − α2β 2 = αβ 2

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Probability Density Function (α = k, β =


θ)

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Cumulative Distribution Function (α =


k, β = θ)

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

α and β both play a role in


determining the mean and the
variance of the random variable
Curves are not symmetric and
are located entirely to the right
of the vertical axis
For α > 1, the maximum value of
the density occurs at the point
x = (α − 1)β
Chandra Shekhar MATH F113 (Probability and Statistics)
Gamma Distribution (Cont...)

Example 4.3/pp.143 Let X be


a gamma random variable with pa-
rameters α = 3 and β = 4
(a) What is the probability density
function?
−x
(
1 α−1 β
Γ(α)β α x e ; x>0
f (x) =
0 ; elsewhere

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

Hence,
1 2 − x4

f (x) = 128 x e ; x>0
0 ; elsewhere

Chandra Shekhar MATH F113 (Probability and Statistics)


Gamma Distribution (Cont...)

(b) What is the moment generating


function for X
1 −3 1
mx(t) = = (1 − 4t) , t < .
(1 − βt)α 4
(c) Find mean, variance and stan-
dard deviation
µ = αβ = 12
σ 2 = αβ 2 = 48
√
σ = 48 = 6.9282

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution

Exponential Distribution
In Gamma Distribution, Put α = 1,
we get
( −x
1
βe x > 0, β > 0
β
f (x) =
0 elsewhere
or if λ = 1/β,
(
λe−λx x > 0, λ > 0
f (x) =
0 elsewhere
Chandra Shekhar MATH F113 (Probability and Statistics)
Exponential Distribution (Cont...)

Exponential Distribution PDF

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

The c.d.f of exponential distribution


with parameter β is given by
Z x Z x
1 − βt
= f (t)dt = e dt
0 0 β
x
1 1 − βt x
= e = 1 − e− β
β (−1/β) 0
Thus,
( x
1 − e− β ; x > 0
F (x) =
0 ; otherwise
Chandra Shekhar MATH F113 (Probability and Statistics)
Exponential Distribution (Cont...)

Exponential Distribution CDF

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

Moment generating function, Mean


and Variance of exponential distri-
bution
Note: Put α = 1 in the gamma distri-
bution, we get the required results.
1
mx(t) = (1 − βt)−1 t <
β
E [X] = β
V ar(X) = β 2

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

The distribution arises in practice in


conjunction with the study of Pois-
son processes, where we have dis-
crete events are being observed in
continuous time interval. If we let
W denote the time of the occurrence
of the first event, then W is a con-
tinuous random variable

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

Theorem
Consider a Poisson process with
parameter λ. Let W denote the time of
the occurrence of the first event. W has
an Exponential distribution with
1
β=
λ

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

Proof This theorem is distribution


of waiting time. The distribution func-
tion F for W is given by
F (w) = P [W ≤ w] = 1 − P [W > w]
Here, we note that, the first occur-
rence of the event will take place af-
ter time w only if no occurrence of
the event are recorded in the time
interval [0, w)
Chandra Shekhar MATH F113 (Probability and Statistics)
Exponential Distribution (Cont...)

Let X denote the number of occur-


rences of the event in this time in-
terval [0, w).
X is a Poisson random variable with
parameter λw. Thus,
P [W > w] = P [X = 0]
e−λw (λw)0
= = e−λw
0!

Chandra Shekhar MATH F113 (Probability and Statistics)


Exponential Distribution (Cont...)

By substitution, we get
F (w) = 1 − P [W > w] = 1 − e−λw
Since, in the continuous case, the deriva-
tive of the cumulative distribution
function is the density
F 0(w) = f (w) = λe−λw
This is exactly density for an expo-
nential random variable with β = λ1
Chandra Shekhar MATH F113 (Probability and Statistics)

You might also like