You are on page 1of 7

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/283711673

A Class of Weibull Mixture Distributions

Article  in  Journal of Biometrics & Biostatistics · December 2012


DOI: 10.4172/2155-6180.1000135

CITATION READS

1 24

1 author:

Mian Adnan
Indiana University Bloomington
48 PUBLICATIONS   43 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Master's Thesis View project

All content following this page was uploaded by Mian Adnan on 18 September 2017.

The user has requested enhancement of the downloaded file.


Adnan and Kiser, J Biomet Biostat 2012, 3:2
Biometrics & Biostatistics http://dx.doi.org/10.4172/2155-6180.1000135

Research
Review Article
Article Open
OpenAccess
Access

A Class of Weibull Mixtured Distributions


Mian Arif Shams Adnan* and Humayun Kiser
Department Of Statistics, Jahangirnagar University, Savar, Dhaka 1342, Bangladesh

Abstract
We have derived a class of mixture distributions which we call weibull mixtures of distributions. Estimation of
unknown parameters along with some properties of these distributions are also prescribed.

Keywords: Mixing distribution; Mixtured distribution; Weibull Definition 3.2


distribution
If X follows a weibull mixture of Normal distribution with
Introduction parameters a and b, then the density function is given by
Mixture distribution was first coined in 1894. A number of authors 1
− x2
like Pearson, Rider, Blichke, Chahine, Roy, et al., authors [1-13]
∞ e 2
x2r
f ( x; a, b) = ∫ abr e b −1 − ar b
1
dr ; −∞ < x < ∞ (3.2)
defined mixtures of two distributions and studied various mixtured 0 r+  1
2 r + 
2
distributions which they called poisson mixture, binomial mixture,  2
negative binomial mixture, Chi-square mixture, Erlang mixture, with parameters a and b such that
Laplace mixture, Rayleigh mixture, F, Dual mixture of distributions.
Weibull distribution is widely being used in bio-statistics, but weibull ∞

mixture distribution has not yet been premeditated. In the present ∫ −∞


f ( x; a, b)dx = 1. (3.3)
paper, we define first the weibull mixture of distributions and then The characteristic function and moments of the same distribution
weibull mixtures of normal, lognormal, gamma, exponential, beta, are presented in the theorem below.
rectangular, erlang, chi-square, t and F distributions and studied some
of their properties. Theorem 3.1
Preliminaries If X has a weibull mixture of normal distributions with parameters
a and b then its characteristic function is represented as
A mixture distribution is a weighted average of probability
1
distribution of positive weights that sum to one. The distributions thus ∞ e
− t2
2  2r  1
2m r + −m  1 
∑   ( it ) 2
r
mixed are called the components of the mixture. The weights themselves φx (t ) = ∫ abr eb −1 − ar b 2
 r + − m dr (3.4)
0 r+
1
 1
m =0
2
 m  2 
comprise a probability distribution called the mixing distribution. 2 r + 
2

 2
Because of this property of weights, a mixture is in particular again  1 
a probability distribution. Mixtures occur most commonly when the r + + s
∞  2 
parameter θ of a family of distributions, given by the density by the and the 2s moment about origin is
th

b −1 − ar b
abr e 2 s
dr
0
 1
density function f(x, θ), is itself subject to the change variation. The r + 
 2
mixing distribution g(x; θ)is then a probability distribution on the
parameter of the distributions. The general formula for the finite and (2s+1)th moment about origin is zero. Mean = 0,
mixture is ∑ i =1 f ( x;θi )g (θi ) ; the infinite analogue, in which g is a density
k

function, is ∫ f ( x;θ ) g (θ )dθ .


 −
1
 1 −
2
 2 
Main Results −
1
 1 
3 + 8a b 1 +  + 4a b 1 +  
 b  b  
Varinace = 1 + 2a 1 +  , β1 = 0, β 2 = 
b
.
Here in this paper, we define the weibull mixtures of some well  b  −
1
 1 
2

1 + 2a b 1 +  
known distributions such as normal, lognormal, gamma, exponential,   b  
beta, rectangular, erlang, chi-square, t and F distributions. Then some
characteristics of these distributions such as characteristic functions, Remark: For a = b = 0, φx (t ), µ2′ s , µ(2′ s +1) , µ1 , µ2 , µ3 , µ4 , β1 and β2 are
moments, and shape chaacteristics are also obtained. The main results same for Normal distribution with mean zero and variance unity.
of the paper are presented in form of definitions and theorems.
Comparison of the probability density functions and the first two
moments are prescribed in the tertiary section.
Definition 3.1 *Corresponding author: Mian Arif Shams Adnan, Department Of Statistics,
Jahangirnagar University, Savar, Dhaka 1342, Bangladesh, E-mail:
A random variable X is said to have a weibull mixtured distribution julias284@yahoo.com
if its probability density function is defined as
∞ Received March 06, 2012; Accepted March 24, 2012; Published March 24, 2012
f ( x; a, b, α ) = ∫ abr b −1e − ar g ( x; α )dr
b

0
(3.1) Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J
Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135
Where g(x,α) is a probability density function. The name of weibull
mixture distribution comes from the fact that the distribution (3.1) is Copyright: © 2012 Adnan MAS, et al. This is an open-access article distributed
the weighted average of g(x,α) with weights equal to the ordinates of under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the
weibull distribution. original author and source are credited.

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135
Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135

Page 2 of 6

Definition 3.3  1 2 3 
 2 − − − 
b 1 + 1  + (6α + 11)a b 1 + 2  + 6a b 1 + 3 
If a random variable X has the density function 3α + 6α + (6α + 6)a       
  b  b  b 
1
− (log x )2  2 
∞ e 2
(log x) 2 r  −4  1   1  2 
f ( x; a, b) = ∫ abr e b −1 − ar b
dr ; x > 0 (3.5)  
b 1+ 4  
 − (6α + 8) a
− 1   −
b 1 +  − 12 a b 1 +  a b 1 + 
   1    −
 2  
1 +a        
0 r+  1   b   b    b   b  
x2 2
r +       
 2 
 1   3   1 
3 
 
then it is said to have a weibull mixture of Lognormal distribution with − − −
 −4 a b 1 + 1   a b 1 + 3   + 6 a b 1 + 1   
         
parameters a,b since  
 b    b     b   
      

∫ 0
f ( x; a, b)dx = 1 (3.6)   1
  −
1
2
  2
  − 2
  1
 
 +6 a b 1 +   a b 1 +   − 3 a b 1 +  
   


1



4 


   b    b    b  
Various moments of the distribution are given in the next theorem.        
β = .
2 2
Theorem 3.2 
 −
1

2 

1 
2

α + a b 1 + 1  + a b 1 + 2  − a b 1 + 1   
If X is a weibull mixture of lognormal variable with parameters a,b  
 b
 
 b 
  
 b  
  
then its characteristic function is given by  
∞ 1 (it ) k 12 k 2 Remark: ϕx(t),μ1,μ2,μ3,μ4,β1 and β2 are true for Gamma distribution
φx (t ) = ∫ abr b −1e − ar
b



1 e with parameters α and β when a = b = 0. For α = 1, weibull mixture
0 r+  1 k =0
k!
2 r +  of Gamma distribution should be equivalent to weibull mixture of
2

 2 Exponential distribution. As such we also derived the weibull mixture


of Exponential distribution.
 2r  2 r − 2 m m + 12  1
∑ m=0  2m  k 2  m + 2 dr
r
(3.7) Estimates of parameters by the method of moments: Let X1, X2,
  X3,……..,Xm be a random sample from the distribution (3.8). We assume
and the sth moment about origin is that parameters a,b and β are known. Then the distribution contains
1 1 
1
1 2 − 
e 2
s
 2r  2 r − 2 m  1
only one unknown parameter α. We have µ1′ = β α + a b 1 + b   ,

   

r

b
b −1 − ar
abr e s  m + dr .
and m1′ = ∑ i = X . Hence by the method of moments, we get,
m =0  x
0
r −m  1  2m   2
2 r +  m
 2 1
1 −
1
b 
1  −
b 
1
Definition 3.4 α + a 1 +   = X . Therefore, αˆ = X β − a  1 +  . (3.11)
β   b    b
A random variable X having the density function Definition 3.5
α + r − β x α + r −1
∞ b β e x
f ( x; a, b, α , β ) = ∫ abr b −1e − ar dr ; x > 0 (3.8) A random variable X having the density function
0
(α + r ) ∞ α r +1e −α x x r
f ( x; a, b, α ) = ∫ abr b −1e − ar
b
is defined a weibull mixture of Gamma distribution with parameters dr ; x > 0 (3.12)
a,b,α and β whereas
0
( r + 1)

is said to have a weibull mixture of Exponential distribution with
∫0
f ( x; a, b, α , β )dx = 1. (3.9)
parameters a,b α, and
The characteristic function and moments are followed by the next ∞
theorem. ∫0
f ( x; a, b, α )dx = 1 (3.13)
Theorem 3.3 Theorem 3.4
If X denotes a weibull mixture of gamma variate with parameters If X follows weibull mixture of exponential distributions with
a,b,α and β then its characteristic function is obtain as parameters a,b and α then its characteristic function is given by
−α  it 
 it  − ar b − r ln 1− 
it 
−1  it 
dr (3.10) − ar b − r ln 1− 


∫  β ∞
φx (t ) = ab 1 − r b −1e dr , (3.14)
 β  0
φx (t ) = ab 1 −
 α  ∫
0
r b −1e  α

1  1  1  1 
1 1
− −
Mean = α + a b 1 +   , Mean = 1 + a b
1 +   ,
β   b   α   b  
 2
  2

 2   − b  1   
1 2 1
1  −
1
b 
1 −
2
b 
2   − b1  1    1 −  1 −
Variance = 2 1 + a b 1 +  + a b
Variance = 2 α + a 1 +  + a 1 +  − a 1 +   , 1 +  − a  1 +  ,
β   b  b    b    α   b  b    b   
   
2

2
  − b  1   
2

 3   − b  1   
2 1 2 3 1
1
1
2
2
3 1
 1 b 
2 b 
3
 2α + 2a b  1 +  + 3a b  1 +  + a b
− − − − − −

1 +  − 3 a 1 +   −  2 + 2a b  1 +  + 3a  1 +  + a  1 +  − 3  a  1 +  
  b  b  b    b      b  b  b   b   
 3   3

  − b  1    − b  2    − b1
1 2
 1 1      − b1  1    − b  2    − b  1  
2 1

3  a  1 +    a  1 +   + 2  a  +    −3 a  1 +  a  1 + +
  2 a  1 +  
   b     b     b       b     b     b    ,
β1 = 3
, β1 = 3
 1 2
 − 1 1   
2
 1
1
2
2 

1
1 

2

1 + a b  1 +  + a b  1 +  − a b  1 +   
1 2
α + a b  1 +  + a b  1 +  − a b  1 +   
− − − − −

  b  b    b      b  b    b   

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135
Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135

Page 3 of 6

 1 2 3   1 2 
 − − −   2 − − 
b 1 + 1  + 17a b 1 + 2  + 6a b 1 + 3  \ b 1 + 1  + (6α + 11)a b 1 + 2 
9 + 12a        3α + 6α + (6α + 6)a     
  b   b   b     b  b 
 2   2 
 −4  1   1  2   3 4  1  
− − − − 3 − 4  − 1 
 
b 1+ 4   1  
b 1 +  − 12 a b 1 +  a b 1 + 
   1     2    
b 1+  b 1+  
b 1+   
+a   − 14 a         +6a  +a   − (6α + 8) a   
 b   b    b   b    b   b    b 
  
         
 3    1  2   1  3  
  −1  3

  1

    −
 1 

 2   −
 1 

 3  
 −4 a b 1 + 1   a b 1 + 3   + 6 a b 1 + 1     −12 a b 1 +   a b 1 +   − 4 a b 1 +   a b 1 +   
             b    b     b    b   
 b   b    b 
            
    
 
  1 3 2 4
 2 4    1   2   1  
  −
1   2

  1

    −  1   −  1   −  2   −  1  
 +6  a b  1 + 1    a b  1 + 2   − 3  a b  1 + 1     +6 a b 1 +   + 6 a b 1 +   a b 1 +   − 3 a b 1 +   
         b     b    b    b  
   b    b    b  
        
        . β = .
β = 2
 2 2
2
 2 2  −
1

2 

1  
1 2  1  
 − − − α + a b 1 + 1  + a b 1 + 2  − a b 1 + 1   
1 + a b  1 + 1  + a b  1 + 2  −  a b  1 + 1     
 b
 
 b 
  
 b  
      
 b  b   b      
    
 
Remark: When a = b = 0, then all of ∏x(t),μ1,μ2,μ3,μ4,β1 and β2 are Remark: a = b = 0 provides all the values of ϕx(t),μ1,μ2,μ3,μ4,β1 and
similar to these of Exponential distribution with parameter α. β2 to be true for Erlang distribution with parameters α and β.
Parameter estimation: If X1, X2, X3,……..,Xm be a random Estimating parameters: For a random sample X1, X2, X3,……..,Xm
sample drawn from the distribution (3.12) and parameters a,b are from the distribution (3.16), we assume that parameters a,b and β are
assumed known, then the distribution contains only one unknown 1 
1
1 
known and α unknown parameter. Here, µ1′ = α + a b 1 +   , and

parameter α. So,
1 −
1
 1 
µ1′ = 1 + a b 1 +   , and m1′ = ∑ xi = X . Therefore, αβ   b  
α   b   m ∑x = X . We obtain
1  −
1
 1 
1 +   = X . Therefore,
i
m′ = α + a b
 −
1
 1 
1
m αβ   b  
1 −
1
 1  1 + a b 1 +  
1 + a 1 +   = X . Hence,  b   (3.15)
b
 1
α   b   αˆ =   1
a

1 + 
b
X
 b
Definition 3.6 αˆ = (3.19)
( X β − 1)
If a random variable X has the density function
∞ (αβ )α + r e −αβ x xα + r −1 Definition 3.7
f ( x; a, b, α , β ) = ∫ abr b −1e − ar
b
dr ; x > 0 (3.16)
0
(α + r ) A random variable X having the density function
then it is said to have a weibull mixture of Erlang distribution with ∞ (r + 1) x r
f ( x; a, b, m) = ∫ abr b −1e − ar
b
dr ;0 < x < m (3.20)
parameters a,b,α and β since 0 m r +1
∞ is said as weibull mixture of Rectangular distribution with parameters
∫0
f ( x; a, b, α , β )dx = 1 (3.17) a,b and m satisfying
The characteristic function as well as the moments is stated in the m

following theorem. ∫0
f ( x; a, b, m)dx = 1. (3.21)
Different moments of the above mentioned distribution are
Theorem 3.5 expressed below.
If X has weibull mixture of erlang distributions with parameters
Theorem 3.6
a,b, α and β then its characteristic function is given by
−α  it  If X follows a weibull mixture of rectangular distribution with
 it  ∞ − ar b − r ln 1− 
parameters a,b and m then its characteristic function is obtained as
φx (t ) = ab 1 − ∫  αβ 
b −1
 r e dr (3.18)
 αβ  0
∞ (it ) k (r + 1)m r + k +1


φx (t ) = ∫ abr b −1e − ar
b
dr (3.22)
1   1 
1
k !m r +1 (r + k + 1)
− k =0
0
Mean = α + a b 1 +   ,
αβ   b   ∞ r +1
and the sth moment about origin is m s ∫ abr b −1e − ar
b
dr .
 2
 0 r + s +1
 2   −  1  
1 2 1
1  −  1 −
Variance = α + a b 1 +  + a b 1 +  − a b 1 +    , Remark: If a = b = 0 then all the values of ϕx(t),μ1,μ2,μ3,μ4,β1 are true
(αβ ) 
2
 b  b    b   
  for Rectangular distribution with parameter m.
2
 1
1
2
2
3
3  − 1 1   
2

 2α + 2a b  1 +  + 3a b  1 +  + a b  1 +  − 3 a b  1 +    Definition 3.8
− − −

  b  b  b   b   
 3
 A random variable X having the density function
  − b1  1    − b2  2    − b1  1    ∞ xα + r −1 (1 − x) β −1
 −3 a  1 +    a  1 +   + 2  a  1 +    f ( x; a, b, α , β ) = ∫ abr b −1e − ar
b
dr ;0 < x < 1 (3.23)
   b     b     b    , 0 B(α + r , β )
β1 =
2 3
 1
1
2
2  −
1

1   is said to have a weibull mixture of Beta distribution of 1st kind with
α + a b  1 +  + a b  1 +  − a b  1 +   
− −

  b  b    b    parameters a,b, α and β. Here we have

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135
Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135

Page 4 of 6

1 Definition 3.10
∫ 0
f ( x; a, b, α , β )dx = 1 (3.24)
A random variable X2 with the density function
Theorem 3.7
1 n
− χ2 + r −1
If X follows weibull mixture of beta distributions of first kind with ∞ e 2
(χ 2 ) 2
f ( χ 2 ; a, b, n) = ∫ abr b −1e − ar
b

n
dr ; χ 2 > 0 (3.28)
parameters a,b, α and β, then its sth moment about origin is given by
0 +r n
2 2
+r
∞ B(α + s + r , β ) 2
φx (t ) = ∫ abr b −1e − ar
b
dr (3.25)
0 B(α + r , β ) is said to have a weibull mixture of Chi-square distribution having the
Remark: For a = b = 0, all the values of µ s′ , µ1′, µ2′ and μ2 are true for parameters a,b and n since
Beta distribution of 1st kind with parameters α and β. ∞

Definition 3.9
∫ 0
f ( χ 2 ; a, b, n)d χ 2 = 1. (3.29)
Theorem 3.9
A random variable X having the density function
∞ x α + r −1 If X2 has weibull mixture chi-square distribution with parameters
f ( x; a, b, α , β ) = ∫ abr b −1e − ar
b
dr ; x > 0 (3.26)
0 B(α + r , β )(1 + x) α +β +r
a,b and n then its characteristic function is expressed as
is called a weibull mixture of Beta distribution of 2nd kind with −
n

∫ dr (3.30)
b

parameters a,b, α and β. Moreover, φx (t ) = ab(1 − 2it ) 2


abr b −1e − ar − r ln(1− 2 it )
0

∫ 0
f ( x; a, b, α , β )dx = 1 (3.27)
Mean = n + 2a

1
 1
1 +  ,
b
Next theorem presents some properties of the same distribution.  b

Theorem 3.8 Remark: Setting a=b=0 we find that all the values of
ϕx(t),μ1,μ2,μ3,μ4,β1 and β2 are true for Chi-square distribution with
If X follows weibull mixture of beta distribution of second kind with parameters n.
parameters a,b, α and β then its sth moment about origin is given by Parameter estimation: Let X1, X2, X3,……..,Xm be a random sample
∞ B(α + s + r , β − s ) from the distribution (3.28). We assume that parameters a and b are

b
abr b −1e − ar dr .
B(α + r , β )
and m′ = ∑ xi = X .
0 1
−  1
known and n is unknown. Now, µ1′ = n + 2a b
1 +  ,
 b 1
m
Remark: Putting a = b = 0 then all the values of µ s′ , µ1′, µ2′ and µ2
1 1
As such, n + 2a
−  1 −  1  (3.31)
are true for Beta distribution of 2 kind with parameters α and β. 1 +  = X . Therefore, nˆ = X − 2a
nd
1 + 
b b
 b   b

Sl. Name of the distribution Probability density function f(x) Support Parameters
1
− x2
2r
∞ e 2
x

b
abr b −1e − ar dr
1 Weibull mixtured Normal 0 r+
1
 1 -∞ < x < ∞ a,b
2 r + 
2

 2
1
− (log x )2
∞ e 2
(log x) 2 r

b
abr b −1e − ar dr
2 Weibull mixtured Lognormal 0 r+
1
 1 x>0 a,b
x2 2
r + 
 2

∞ β α + r e − β x xα + r −1

b

3 Weibull mixtured Gamma abr b −1e − ar dr x>0 a,b,α,β


0
(α + r )
∞ α r +1e −α x x r

b

4 Weibull mixtured Exponential abr b −1e − ar dr x>0 a,b,α


0
( r + 1)
∞ (αβ )α + r e −αβ x xα + r −1

b

5 Weibull mixtured Erlang abr b −1e − ar dr x>0 a,b,α,β


0
(α + r )
r
∞ (r + 1) x

b
6 Weibull mixtured Rectangular abr b −1e − ar dr 0<x<m a,b,m
0 m r +1
∞ xα + r −1 (1 − x) β −1

b

7 Weibull mixtured Beta 1st kind 0


abr b −1e − ar
B(α + r , β )
dr 0<x<1 a,b,α,β
α + r −1
∞ x

b

8 Weibull mixtured Beta 2nd kind 0


abr b −1e − ar
B(α + r , β )(1 + x)α + β + r
dr x>0 a,b,α,β
1 n
− χ2 + r −1
∞ e 2
(χ ) 2 2


b
abr b −1e − ar dr
9 Weibull mixtured Chi-square 0 n
+r n χ2 > 0 a,b,n
22 +r
2
∞ t 2r

b
abr b −1e − ar n +1
dr
0
10 Weibull mixtured t­ 1
+r 1 n  t2  2
+r
-∞ < t < ∞ a,b,n
n 2
B  + r ,  1 + 
2 2  n
n1
+r n1
 n1  2 + r −1
  F2

 n2 

b −1 − ar b
11 Weibull mixtured F 0
abr e n1 + n2
+r
dr F>0 a,b,n1,n2
n n  n  2
B  1 + r , 2  1 + 1 F 
2 2   n2 

Table 1: Comparison of density functions of different Weibull mixture distributions. χ2>0.

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135
Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135

Page 5 of 6

Sl. Name of the distribution Mean Variance

1
−  1
1 Weibull mixtured Normal 0 1 + 2a b
1 + 
 b
2 Weibull mixtured lognormal can be obtained from equation 3.7 can be obtained from equation 3.7
 −
1
 1 −
2
 2 
α + a b 1 +  + a b 1 +  
1 −
1
 1  1   b   b 
3 Weibull mixtured Gamma α + a b
1 +   
β 2   − 1  1  2

β   b   
 − a 1 +  
b

   b   

 1
 1
2
2 
b 
− −
1 + a b 1 +  + a 1 +  
1 −
1
 1  1   b  b 
4 Weibull mixtured Exponential 1 + a b 1 +    2 
α   b   α 2   − 1  1   
 − a 1 +
b
  
   b   

 −
1
 1 −
2
 2 
α + a b 1 +  + a b 1 +  
1  −
1
 1  1   b  b 
5 Weibull mixtured Erlang α + a b 1 +    2 
αβ   b   (αβ ) 2   − b  1  
1

 − a 1 +   
   b   
6 Weibull mixtured Rectangular can be achieved from equation 3.22 can be achieved from equation 3.22
7 Weibull mixtured Beta 1st kind equation 3.25 provides equation 3.25 provides

1  −
1
 1 −
2
 2 
α (α + 1) + (2α + 1)a b 1 +  + a b 1 +  
( β − 1)( β − 2)   b  b  
1  −
1
 1 
8 Weibull mixtured Beta 2nd kind α + a b 1 +   2
β − 1   b    1  −
1
 1  
− α + a b 1 +  
 β − 1   b  

 −
1
 1 −
2
 2 
 2n + 4a b  1 +  + 4a b  1 +  

1
 1   b  b 
9 Weibull mixtured Chi-square n + 2a b
1 +   2 
 b   −  1  
1

 −4 a 1 +  
b

   b   

n  −
1
 1 
10 Weibull mixtured t 0 1 + 2a b 1 +  
n − 2   b  

 −
1
 1 
 n1 (n1 + 2) + 4(n1 + 1)a b 1 +  
n22
  b 
2
n1 (n2 − 2)(n2 − 4)  2 
n2  −
1
 1   −  2 
 n1 + 2a b +4a b 1 + 
11 Weibull mixtured F 1 +     b  
n1 (n2 − 2)   b    
2
 n2  −
1
 1  
− n1 + 2a b 1 +  
 n1 (n2 − 2)   b  

Table 2: Comparison among first two moments of different Weibull mixtured distributions.
Definition 3.11 The following theorem expresses here some of the properties of the
distribution.
If t as a random variable has the density function

f (t ; a, b, n) = ∫ abr b −1e − ar
b t 2r
dr ; −∞ < t < ∞
Theorem 3.10
0 n +1
1 +r
1 n  t2  2 +r If t is weibull mixture of t distribution with parameters
n B  + r ,  1 +  2

 2 2  n (3.32)
then it is said to have a weibull mixture of t distribution with parameters a,b and n then the 2sth moment about origin is given by
a,b and n if  1 n 
r + s +   − s
∞  2 2 
∞ n s

b −1 − ar b
abr e dr and the (2s+1)th moment about origin
∫−∞
f (t ; a, b, n)dt = 1 (3.33) 0
1  n
 +r  
2  2

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135
Citation: Adnan MAS, Kiser H (2012) A Class of Weibull Mixtured Distributions. J Biomet Biostat 3:135. doi:10.4172/2155-6180.1000135

Page 6 of 6

 1
1
2
2 
Remark: For a = b = 0 all the values of ϕx(t),μ1,μ2 and μ2 are true
b  b 
− −
 3 + 8a  1 +  + 4 a  1 +   for F distribution with parameters n1 and n2.
n−2   b   b 
is zero, β1 = 0, β 2 = .
n−4  −
1
b 
1  
2
Comparison
1 + 2a  1 +  
  b  A Comparison among various features of the different weibull
Remark: If a = b = 0 then all the values of μ2s+1, μ2s, μ1,μ2,μ3,μ4,β1 and mixtured distributions is shown in the following table 1 and table 2.
β2 are true for t distribution with parameter n. References
Definition 3.12 1. Pearson K (1894) Contributions To The Mathematical Theory of Evolution.
Philos Trans R Soc London 185: 71-110.
A random variable F having the density function
n1 2. Rider PR (1961) The Method of Moments Applied To A Mixture of Two
+r
 n1  2 n1
+ r −1 Exponential Distributions. The Annals of Mathematical Statistics 32: 143-147.
  F 2

f ( F ; a, b, n1 , n2 ) = ∫ abr b −1e − ar
b
 n2  dr ; F > 0
3. Blichke WR (1965) Mixtures of Discrete Distribution. Classical And Contagious
0 n1 + n2 Discrete Distributions. Calcutta Statistical Publishing Society. Ed GP Patil 351-
+r
n n  n  2 372.
B  1 + r , 2  1 + 1 F 
2 2   n2  4. Chahine J (1965) Une Generalization Dela Loi Binomiale Negative. Rev Stat
(3.34) Appl 13: 33-43.

is said to have a weibull mixture of F distribution with parameters a,b, 5. Roy MK, Rahman S, Ali MM (1992) A Class Of Poisson Mixtured Distributions.
J Info And Optimiz Sci 13: 207-208.
n1 and n2 if
6. Roy MK (1998) Some Negative Binomial Mixtured Distributions. J Info And

∫ 0
f ( F ; a, b, n1 , n2 )dF = 1 (3.35) Optimiz Sci 19: 285-302.

7. Roy MK, Sinha SK (1995) Negative Binomial Mixture of Normal Moment


The following theorem presents the characteristic function and Distributions. Metron 3: 83-92.
moments of this distribution. 8. Authors (2009) Some Laplace Mixture Distributions. J Appli Stati Sci 17: 549-
560.
Theorem 3.11
9. Authors (2010) Some Pareto Mixtured Distributions. Proceedings. American
If F follows weibull mixture of F distribution with parameters a,b, Statistical Association.
n1 and n2 then its characteristic function is given by 10. Authors (2011) A Class of Rayleigh Mixture Distributions. Journal of
x Interdisciplinary Mathematics 14: 507-522.
 n2   n2   n2 
 it   + r + x  − x 11. Author (2011) A Mixture Probabilistic Model for Extreme Distributions.
n  2   2 
φx (t ) = ∫ abr b −1e − ar ∑ x =0  1 
∞ b ∞
dr Proceedings. American Statistical Association.
 n1   n2 
0 x! 12. Authors (2012) A Class of Mixture Distributions. Journal of Interdisciplinary
 +r  
2   2 Mathematics. To appear.
(3.36) 13. Authors (2012) A Class of Dual Mixture Distributions. Journal of Interdisciplinary
 n2   n2  Mathematics. To appear.
s  + r + s  − s
 2   2
and the Sth moment about origin is  n2   14. Johnson NL, Kotz S, Kemp AW (1992) Univariate Discrete Distributions. (2nd


b
abr b −1e − ar dr ,
 n1  0
 n1   n2  Edition) Wiley, New York.
 +r  
2   2 15. Johnson NL, Kotz S, Balakrishnan N (1995) Continuous Univariate Distributions.
(2nd Edition) Wiley, New York.

Submit your next manuscript and get advantages of OMICS


Group submissions
Unique features:

• User friendly/feasible website-translation of your paper to 50 world’s leading languages


• Audio Version of published paper
• Digital articles to share and explore
Special features:

• 200 Open Access Journals


• 15,000 editorial team
• 21 days rapid review process
• Quality and quick editorial, review and publication processing
• Indexing at PubMed (partial), Scopus, DOAJ, EBSCO, Index Copernicus and Google Scholar etc
• Sharing Option: Social Networking Enabled
• Authors, Reviewers and Editors rewarded with online Scientific Credits
• Better discount for your subsequent articles
Submit your manuscript at: http://www.omicsonline.org/submission

J Biomet Biostat
ISSN:2155-6180 JBMBS, an open access journal Volume 3 • Issue 2 • 1000135

View publication stats

You might also like