Professional Documents
Culture Documents
Lecture 4
Lecture 4
January 3, 2012
Denition 1.1. The Exponential Distribution: An Extremely useful distribution, particularly in internet trac, insurance, catastrophe and rainfall modeling, failure or reliability problem.
f (x) =
0 if x < 0 ex if x 0
The positive number is called the rate of the distribution the corresponding c.d.f is given by the following formula :
F (x) =
0 if x < 0 (1 ex ) if x 0
Denition 1.2. The Cauchy Distribution: The density function can be given by fm, (x) = 1 , xR [2 + (x m)2 ]
Fm, (x) =
fm, (x)dx = 1
1 (x m) 1 tan1 + 2
[2
|x|dx + (x m)2 ]2
does not exist. Therefore, the higher order moments also do not exist. Tail of cauchy is heavier than normal. Denition 1.3. Weibull Distribution: The probability density function can be written as fW E (x; , ) = x1 e(x) ; 0 otherwise
Here, is scale parameter and determines the shape parameter. For < 1, shape parameter of weibull decays sharply like exponential distribution whereas > 1 we nd Weibull as bell-shaped. For = 2 we get Rayleigh distribution. Denition 1.4. Beta Distribution: This distribution is dened on [0,1]. The p.d.f. can be given by f (x; 1 , 2 ) = where B(1 , 2 ) = 1 2 1 + 2
1 x1 1 (1 B(1 ,2 )
x)2 1 ;
0<x<1 otherwise
More on Discrete distribution: We have already mentioned about Geometric distribution. Lets calculate the mean and variance for the distribution in some dierent way. If we consider the version of p.m.f. as P [X = x] = pq x ; x = 0, 1,
Then calculation of E(X) and V (X) can be done as follows : Start with an identity : 2
qx =
x=0
1 1q
xq x1 =
x=0
1 (1 q)2
x=0
xpq x = q p
q p
E(X) =
x(x 1)q x2 =
x=1
2 (1 q)3
x=1
x(x 1)q x =
2q 2 p2 2q 2 p2
Observe: V (X) = + 2 ; variance > mean for geometric distribution. Now lets come to the more general version of Geometric distribution i.e. Negative Binomial. = mean