Hindawi Publishing Corporation

Journal of Applied Mathematics

Volume 2011, Article ID 238290, 17 pages
doi:10.1155/2011/238290

Research Article
Rayleigh Mixture Distribution

Rezaul Karim, Pear Hossain, Sultana Begum, and Forhad Hossain

Department of Statistics, Jahangirnagar University, Savar, Dhaka 1342, Bangladesh

Correspondence should be addressed to Rezaul Karim, [email protected]

Received 4 June 2011; Accepted 2 October 2011

Academic Editor: Tak-Wah Lam

Copyright q 2011 Rezaul Karim et al. This is an open access article distributed under the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

This paper presents Rayleigh mixtures of distributions in which the weight functions are assumed
to be chi-square, t and F sampling distributions. The exact probability density functions of the mix-
ture of two correlated Rayleigh random variables have been derived. Different moments, charac-
teristic functions, shape characteristics, and the estimates of the parameters of the proposed mix-
ture distributions using method of moments have also been provided.

1. Introduction
In statistics, a mixture distribution is expressed as a convex combination of other probability
distributions. It can be used to model a statistical population with subpopulations, where
components of mixture probability densities are the densities of the subpopulations, and the
weights are the proportion of each subpopulation in the overall population. Mixture distri-
bution may suitably be used for certain data set where different subsets of the whole data
set possess different properties that can best be modeled separately. They can be more mathe-
matically manageable, because the individual mixture components are dealt with more nicely
than the overall mixture density. The families of mixture distributions have a wider range of
applications in different fields such as fisheries, agriculture, botany, economics, medicine,
genetics, psychology, paleontoogy, electrophoresis, finance, communication theory, sedimen-
tology/geology, and zoology.
Pearson 1 is considered as the torch bearer in the field of mixtures distributions.
He studied the estimation of the parameters of the mixture of two normal distributions.
After a long period of time, some basic properties of mixture distributions were studied
by Robins 1948. Some of other researchers 2–5 have studied in greater detail the finite
mixture of distributions. Roy et al. 6–12 defined and studied poisson, binomial, negative
binomial, gamma, chi-square and Erlang mixtures of some standard distributions. In the
light of the above-mentioned distributions, here we have studied Rayleigh mixtures of
2 Journal of Applied Mathematics

distributions in which the weight functions are assumed to be chi-square, t- and F-distri-
bution, and the moments, characteristic function, and shape characteristics of these mixtures
distributions have also been studied.

2. Preliminaries
Suppose the random variable X has a probability density function pdf fx | θ and if the
parameter space Θ is a discrete random variable containing parameter values θ1 , θ2 , . . . , θk
such that the distribution of Θ is P Θ  θi   pi , then the unconditional distribution of X is


k
mx  pi fx | θi . 2.1
i1

This is called a mixture of the distributions fx | θi  with weight pi , i  1, 2, . . . , k. The above
definition may be extended to the case for large k.
It can be generalized to the case when the parameter space Θ is absolutely continuous
random variable having pdf τθ. We will have, then, a continuous mixture of densities fx |
θ with weight function τθ. In this case, the unconditional distribution of X is

mx  fx | θτθdθ. 2.2
Θ

3. Main Results
In this paper we first define the general form of Rayleigh mixture distribution. Then we fur-
nished the Rayleigh mixture of some well-known sampling distributions such as chi-square,
t- and F-distributions. The exact distribution of the mixture of two correlated Rayleigh dis-
tributions has been studied.
The main results of this study have been presented in the form of some definitions and
theorems.

Definition 1. A random variable X is said to have Rayleigh mixture distribution if its

probability density function is defined by
   ∞ re−r 2 /2σ 2
f x; σ 2 , n  τx, r; ndr, 3.1
0 σ2

where τx, r; τ, n is a probability density function or any sampling distribution such as chi-
square, t- and F-distribution.
The name Rayleigh mixture distributions is given due to the fact that the derived dis-
tribution 3.1 is the weighted sum of τx, r; τ, n with weight factor equal to the probabilities
of Rayleigh distribution.

3.1. Formulation of Rayleigh Mixture Distribution

The Rayleigh mixtures of distributions in which the weight functions are assumed to be chi-
square, t- and F-distribution. In a statistical theory, we will use chi-square distribution as a
Journal of Applied Mathematics 3

weight function if sampling statistic follows chi-square distribution. For example, sampling
variance is followed by chi-square distribution and we can use chi-square distribution as a
weight function. Similarly, we will use t-distribution and F-distribution if and only if sam-
pling statistic follows t-distribution and F-distribution, respectively. For example, if popu-
lation variance is unknown and sample size is very small, then the sampling mean follows
t-distribution and the ratio of sampling variances follows F-distribution. Now we define Ray-
leigh mixtures of distributions for different weight functions as follows.

3.1.1. Rayleigh Mixtures of Chi-Square Distribution

Definition 2. A random variable χ2 is said to have a Rayleigh mixture of chi-square dis-
tribution with parameter σ 2 with degrees of freedom n if its probability density function
is defined by

   ∞ re−r 2 /2σ 2 e−χ2 /2 χ2 n/2 r−1

f χ ;σ ,n 
2 2
dr; 0 < χ2 < ∞, 3.2
0 σ2 2n/2 r n/2 r

where the weight function τx, r; τ, n in 3.1 is the chi-square sampling distribution. Here
the notation “ ” in 3.2 is a gamma function such that a  a − 1!  a − 1a − 2 · · · 3 · 2 · 1.

3.1.2. Rayleigh Mixtures of t-Distribution

Definition 3. A random variable t is defined to have a Rayleigh mixture of t-distributions with
parameter σ 2 and degrees of freedom n if its probability density function is defined as

   ∞ re−r 2 /2σ 2 t2r

f t; σ 2 , n  dr,
n1/2 r B1/2 r, n/21 t2 /nn 1/2 r
σ 2
0 3.3
−∞ < t < ∞,

where the weight function τx, r; τ, n in 3.1 is the student t-distribution.

3.1.3. Rayleigh Mixtures of F-Distribution

Definition 4. A random variable F is defined to have a Rayleigh mixture of F-distributions
with parameter σ 2 and degrees of freedom n1 and n2 , if its probability density function is
defined as

   ∞ re−r 2 /2σ 2 n1 /n2 n1 /2 r F n1 /2 r−1

f F; σ 2 , n1 , n2  dr,
0 σ2 Bn1 /2 r, n2 /21 n1 /n2 Fn1 n2 /2 r 3.4
0 < F < ∞,

where the weight function τx, r; τ, n in 3.1 is the F-distribution.

4 Journal of Applied Mathematics

3.1.4. Mixture of Two Correlated Rayleigh Distributions

Let X and Y be two independent Rayleigh variables with probability density function pdf.
The joint distribution of X and Y with correlation coefficient ρ−1 ≤ ρ ≤ 1 can be constructed
by the following formula:

      
f x, y  fxg y 1 ρ1 − 2Fx × 1 − 2G y 3.5

which was developed by Farlie-Gumbl-Morgenstern 1979.

Using the formula, the mixture of two correlated distributions is as follows:

  xy
f x, y; σ1 , σ1 ; ρ 
σ1 σ2 2

× e−1/2x /σ1  y /σ2  ρe−1/2x /σ1  y /σ2  − 2ρe−1/22x /σ1  y /σ2 

2 2 2 2 2 2 2 2 2 2 2 2


−2ρe−1/2x /σ12  2y 2 /σ22 
4ρe−x /σ12  y 2 /σ22 
2 2
,
3.6

where x > 0, y > 0; σ1 , σ2 > 0 and −1 ≤ ρ ≤ 1.

3.2. Derivation of Characteristics of Rayleigh Mixture Distribution

Moments and different characteristics of the Rayleigh mixture of distributions are presented
by the following theorems.

Theorem 3.1. If χ2 follows a Rayleigh mixture of chi-square distribution with parameter σ 2 with deg-
rees of freedom n, then the sth raw moment of this mixture distribution about origin is given by

∞
re−r /2σ 2s n/2 r s
2 2

μs  dr. 3.7

0 σ2 n/2 r

Hence, the mean and the variance of this mixture distribution are as follows:

√
Mean  n σ 2π,
√ 3.8
Variance  2n 2σ 2π 2σ 2 4 − π.
Journal of Applied Mathematics 5

Proof. We know the sth raw moment defined by

 s
μs  E χ2

 ∞  n/2 r s−1
re−r /2σ e−χ /2 χ2
2 2 2

 drdχ2 3.9
0 σ2 2n/2 r n/2 r
∞
re−r /2σ 2s n/2 r s
2 2

 dr.
0 σ2 n/2 r

If we put s  1 in 3.9, we get

∞
re−r /2σ 2n/2 r 1
2 2

μ1  dr
σ2 n/2 r
0
3.10
√
 n σ 2π.

If we put s  2 in 3.9, we have

∞
re−r /2σ 22 n/2 r 2
2 2

μ2  dr
0 σ2 n/2 r
∞
re−r /2σ
2 2

4 n/2 r 1n/2 rdr

0 σ2 3.11

√ 3  
 n2 4σn 1 2  2n 8σ 2 On simplification
2
√
 n2 σn 12 2π 2n 8σ 2 .

Hence, the variance is defined by

 2
μ2  μ2 − μ1
√ 3.12
 2n 2σ 2π 2σ 2 4 − π.

This completes the proof.

Theorem 3.2. If χ2 follows a Rayleigh mixture of chi-square distributions with parameter σ 2 and deg-
rees of freedom n, then its characteristic function is given by

∞
re−r /2σ
2 2
−n/2
Φ t  1 − 2it
χ2 1 − 2it−r dr. 3.13
0 σ2
6 Journal of Applied Mathematics

Proof. The characteristic function is defined as

 2
Φχ2 t  E eitχ

 ∞  n/2 r−1
re−r /2σ e−χ /21−2it χ2
2 2 2

 drdχ2
0 σ2 2n/2 r n/2 r
∞ 3.14
re−r /2σ
2 2
1
 2
dr
0 σ 1 − 2itn/2 r
∞
re−r /2σ
2 2
−n/2
 1 − 2it 1 − 2it−r dr
0 σ2

and hence proved

Theorem 3.3. If t follows a Rayleigh mixture of t-distributions with parameter σ 2 and degrees of free-
dom n, then the sth raw moment about origin is

μ2s 1  μ2s 1  0,
∞ 3.15
re−r /2σ r s 1/2 n/2 − s
2 2

μ2s  μ2s  ns dr.

0 σ2 r 1/2 n/2

And hence

n  √
Mean  0, Variance  1 σ 2π for n > 2. 3.16
n − 2

Therefore,

 √
n − 2 2σ 2 2σ 2π 3
Skewness: β1  0, Kurtosis: β2   √ 2 , n > 4. 3.17
n − 4 1 σ 2π

Proof. The 2s 1th raw moment odd order moments about origin is given by


μ2s 1  E t2s 1

∞ ∞
re−r /2σ
2 2
t2r 2s 1
 dr dt
−∞ 0 σ2 n1/2 r B1/2 r, n/21 t2 /nn 1/2 r
∞ ∞
re−r /2σ
2 2
t2r 2s 1
 dt dr
0 σ 2 n1/2 r B1/2 r, n/2 −∞ 1 t2 /n
n 1/2 r
Journal of Applied Mathematics 7
∞ ∞
re−r /2σ
2 2

 2 1/2 r B1/2 r, n/2

ψtdt dr
0 σ n −∞
 
t2r 2s 1
0 Since, ψt  n 1/2 r
is an odd function of t .
1 t2 /n
3.18

If s  0, then μ1  Mean  0. If s  1 then μ3  μ3  0. So, μ2s 1  μ2s 1  0.

Now, the 2sth raw moment about origin is given by

μ2s  μ2s  E t2s
∞ ∞
re−r /2σ
2 2
t2r 2s
 drdt
n1/2 r B1/2 r, n/21 t2 /nn 1/2 r
σ 2
−∞ 0

∞  ∞  2 s r 1/2−1  2 
re−r /2σ ns r 1/2
2 2
t /n t
 d dr
2
0 σ n
1/2 r B1/2 r, n/2 0 1 t /n
2 n 1/2 r n 3.19
∞ ∞  
re−r /2σ ns r 1/2
2 2
us r 1/2−1 t2
 2 1/2 r B1/2 r, n/2 n 1/2 r
dudr; Putting u 
0 σ n 0 1 u n
∞
re−r /2σ r s 1/2 n/2 − s
2 2

n s
dr.
0 σ2 r 1/2 n/2

If s  1 then,

μ2  μ2
∞
re−r /2σ r 3/2 n/2 − 1
2 2

n dr
0 σ2 r 1/2 n/2
  3.20
n 2n √ 3
 2σ  ; n > 2
n−2 n−2 2
 √ 
n  √ 3 π
 1 σ 2π ; ∵   .
n−2 2 2

If s  2, then,
∞
re−r /2σ r 5/2 n/2 − 2
2 2

μ4  μ4  n 2
dr
0 σ2 r 1/2 n/2
 ∞ −r 2 /2σ 2  
n2 re
 r 2 4r 3 dr 3.21
n − 2n − 4 0 σ 2

 
r 5/2 n/2 − 2 r 3/2r 1/2 r 2 4r 3
Since,   .
r 1/2 n/2 n/2 − 1n/2 − 2 n − 2n − 4
8 Journal of Applied Mathematics

Hence,

μ4  μ4
 
n2 √ 3
 2σ 2 4σ 2  3
2
n − 2n − 4 2 3.22

n2  √
 2σ 2 2σ 2π 3 , n > 4.
n − 2n − 4

To find the Skewness and Kurtosis of this mixture distribution

μ23
Skewness: β1   0,
μ32
 √ 3.23
n − 2 2σ 2 2σ 2π 3
μ4
Kurtosis: β2  2   √ 2 .
μ2 n − 4 1 σ 2π

This completes the proof.

Theorem 3.4. If F follows a Rayleigh mixture of F-distributions having parameter σ 2 with degrees of
freedom n1 and n2 , respectively, then the sth raw moment about origin is given by
 s  ∞
re−r /2σ n1 /2 r s n2 /2 − s
2 2
n2
μs  dr. 3.24
n1 0 σ2 n1 /2 r n2 /2

Hence, the mean and variance of this distribution are

n2  √
Mean  n1 σ 2π ,
n1 n2 − 2
⎡  √ 2 ⎤
 2 2 √ 3.25
n2 ⎢ n1 2n1 2n1 1 σ 2π 8σ 2 n σ 2π
1
⎥
Variance  ⎣ − ⎦,
n1 n2 − 2 n2 − 4 n2 − 2 2

respectively.

Proof. The sth raw moment about origin is given by

 ∞
re−r /2σ
2 2
n1 /n2 n1 /2 r F n1 /2 r s−1
μs  EF  
s
drdF
0 σ2 Bn1 /2 r, n2 /21 n1 /n2 Fn1 n2 /2 r
 ∞ −r 2 /2σ 2 ∞
re n1 /n2 n1 /2 r F n1 /2 r s−1
 dFdr 3.26
0 σ2 Bn1 /2 r, n2 /2 0 1 n1 /n2 Fn1 n2 /2 r
 s  ∞ −r 2 /2σ 2
n2 re n1 /2 r s n2 /2 − s
 2
dr.
n1 0 σ n1 /2 r n2 /2
Journal of Applied Mathematics 9

If we put s  1, we get

Mean  μ1
  ∞
re−r /2σ n1 /2 r 1 n2 /2 − 1
2 2
n2
 dr 3.27
n1 0 σ2 n1 /2 r n2 /2
n2  √
 n1 σ 2π .
n1 n2 − 2

Putting s  2, we get

 2  ∞
re−r /2σ n1 /2 r 2 n2 /2 − 2
2 2
n2
μ2  dr
n1 0 σ2 n1 /2 r n2 /2
3.28
 2  √
n2 1
 n21 2n1 2n1 1 σ 2π 8σ 2 .
n1 n2 − 2n2 − 4

Then the variance,

 2
μ2  μ2 − μ1
⎡ √ 2 ⎤

 2 √ 3.29
2
2n1 2n1 1 σ 2π 8σ n
2 σ 2π
n2 ⎢ n1 1
⎥
 ⎣ − ⎦,
n1 n2 − 2n2 − 4 n2 − 2 2

hence proved.

Theorem 3.5. If F follows a Rayleigh mixture of F-distributions having parameter σ 2 with degrees of
freedom n1 and n2 , respectively, then its characteristic function is given by

∞  
re−r /2σ 
2 2 ∞
1 n2 it x n1 /2 r x n2 /2 − x
ΦF t  · dr. 3.30
0 σ 2 x0 x! n1 n1 /2 r n2 /2

From here we may get the mean and variance of this distribution.

Proof. The characteristic function of the random variable F is given by


ΦF t  E eitF

∞ ∞
re−r /2σ
2 2
n1 /n2 n1 /2 r F n1 /2 r−1
 e itF
drdF
0 0 σ2 Bn1 /2 r, n2 /21 n1 /n2 Fn1 n2 /2 r 3.31

∞  
re−r /2σ 
2 2 ∞
1 n2 it x n1 /2 r x n2 /2 − x
 · dr.
0 σ 2 x0 x! n1 n1 /2 r n2 /2
10 Journal of Applied Mathematics

Hence the sth raw moment about origin is

its
μs  coefficient of in ΦF t
s!
 s  ∞ 3.32
re−r /2σ n1 /2 r s n2 /2 − s
2 2
n2
 dr.
n1 0 σ2 n1 /2 r n2 /2

If s  1, then

  ∞
re−r /2σ n1 /2 r 1 n2 /2 − 1
2 2
n2
μ1  dr
n1 0 σ2 n1 /2 r n2 /2
3.33
n2  √
 n1 σ 2π .
n1 n2 − 2

If we put s  2, we get

 2  ∞
re−r /2σ n1 /2 r 2 n2 /2 − 2
2 2
n2
μ2  dr
n1 0 σ2 n1 /2 r n2 /2
3.34
 2  √
n2 1
 n21 2n1 2n1 1 σ 2π 8σ 2 .
n1 n2 − 2n2 − 4

Therefore,

 2
Variance: μ2  μ2 − μ1
⎡ √ 2 ⎤ 
 2 √ 3.35
2
2n1 2n1 1 σ 2π 8σ n σ 2π 2
n2 ⎢ n1 1
⎥
 ⎣ − ⎦.
n1 n2 − 2n2 − 4 n2 − 2 2

Driving coefficient of Skewness  β1 and coefficient of Kurtosis  β2 is a tedious job; we have

avoided the task here.

The different moments of the random variable which is the resultant of the product of
two correlated Rayleigh random variables are obtained by following theorem.

Theorem 3.6. For −1 ≤ ρ ≤ 1, the a, bth product moment of the mixture of two correlated Rayleigh
random variables is denoted by μ a, b; ρ and given by

  a  b     

μ a, b; ρ  σ1a σ2b ×Γ 1 Γ 1 × 2a b/2 ρ 2a/2 − 1 2b/2 − 1 . 3.36
2 2
Journal of Applied Mathematics 11

Proof. We know that

   
μ a, b; ρ  E X a Y b
 ∞
xy
 xa y b ×
0 σ1 σ2 2

× e−1/2x /σ1 y /σ2  ρe−1/2x /σ1 y /σ2  − 2ρe−1/22x /σ1 y /σ2 

2 2 2 2 2 2 2 2 2 2 2 2


−2ρe−1/2x /σ12 2y2 /σ22 
4ρe−x /σ12 y2 /σ22 
2 2
dxdy
 ∞
xa 1 yb 1
× e−1/2x /σ12 y2 /σ22 
2
 2
dxdy
0 σ1 σ2 
3.37
 ∞ a 1 b 1
x y
× e−1/2x /σ12 y2 /σ22 
2
ρ 2
dxdy
0 σ1 σ2 
 ∞
xa 1 yb 1
× e−1/22x /σ12 y2 /σ22 
2
− 2ρ 2
dxdy
0 σ1 σ2 
 ∞
xa 1 yb 1
× e−1/2x /σ12 2y2 /σ22 
2
− 2ρ 2
dxdy
0 σ1 σ2 
 ∞
xa 1 yb 1
× e−x /σ12 y2 /σ22 
2
4ρ 2
dxdy.
0 σ1 σ2 

Now taking the first integral from 3.37

 ∞
xa 1 yb 1
× e−1/2x /σ12 y2 /σ22 
2

2
dxdy 3.38
0 σ1 σ2 

and making a transformation p  1/2x2 /σ12  and q  1/2y2 /σ22  we obtain the following
result:

 √ a  ∞  √ b  ∞ a  b 
σ1 2 pa/2 e−p dp × σ2 2 qb/2 e−q dq  2a b/2 σ1a σ2b × Γ 1 Γ 1 . 3.39
0 0 2 2

Using the similar mathematical simplification we get the following results for the 2nd integral

a  b 
2a b/2 ρσ1a σ2b × Γ 1 Γ 1 , 3.40
2 2

for the 3rd integral

a  b 
2 b/2
ρσ1a σ2b ×Γ 1 Γ 1 , 3.41
2 2
12 Journal of Applied Mathematics

for the 4th integral

a  b 
2a/2
ρσ1a σ2b ×Γ 1 Γ 1 , 3.42
2 2

for the 5th integral

a  b 
ρσ1a σ2b ×Γ 1 Γ 1 . 3.43
2 2

Now putting all of these values in 3.37 and simplifying the result we get our desired result
stated in the theorem.
Special findings of the above theorem if ρ  0 then the product moment of the two cor-
related Rayleigh variables is nothing but the product of ath and bth moments of two inde-
pendent Rayleigh variables. In such case the product moment is as follows:

  a  b 
E X a Y b  2a b/2 × σ1a σ2b × Γ 1 Γ 1
2 2
a   
b 3.44
 2a/2 σ1a Γ 1 × 2b/2 σ2b Γ 1
2 2
 
 EX a  · E Y b .

Theorem 3.7. If X and Y are two correlated Rayleigh variates having joint density given in 3.6,
then probability density function of W  X/Y is given by

  π
f w; σ1 , σ2 ; ρ  × w2 σ1 σ2
2

 √  2 2 −3/2
× 1 ρ 1 2 w σ2 σ12 3.45
 −3/2  −3/2 
−2ρ 2w2 σ22 σ12 w2 σ22 2σ12 ,

where, w > 0; σ1 > 0, σ2 > 0 and −1 ≤ ρ < 1.

Proof. Under the transformation x  z, y  z/w in 3.6 with the Jacobean

   w2
J x, y −→ w, z  , 3.46
z

the pdf of w and z is given by

z2
  2
1 ρ e−z /21/σ1 1/w σ2  − 2ρe−z /22/σ1 1/w σ2 
2 2 2 2 2 2 2
fw, z  2
wσ1 σ2  3.47

−z2 /21/σ12 2/w2 σ22  −z2 1/σ12 1/w2 σ22 
−2ρe 4ρe .
Journal of Applied Mathematics 13

Now integrating over z we get the marginal distribution of w as

  ∞
1 ρ
z2 e−z /21/σ1 1/w σ2 
2 2 2 2
fw  2
dz
wσ1 σ2  0
∞
2ρ
z2 e−z /22/σ1 1/w σ2 
2 2 2 2
− 2
dz
wσ1 σ2  0
∞ 3.48
2ρ 2 −z2 /21/σ12 2/w2 σ22 
− ze dz
wσ1 σ2 2 0
∞
4ρ
z2 e−z 1/σ1 1/w σ2 
2 2 2 2
2
dz.
wσ1 σ2  0

Taking each of the integral separately and making transformation

 −1/2
 1 1
z  2p for the first integral,
σ12 w2 σ22
 −1/2
 2 1
z  2p for the second integral,
σ1 w σ22
2 2

 −1/2 3.49
 1 2
z  2p for the third integral,
σ1 w σ22
2 2

 −1/2
1 1
z p for the fourth integral.
σ1 w σ22
2 2

We have got the following results:

  −3/2   −3/2
π 1 1 π 2 1
, ,
2 σ12 w2 σ22 2 σ12 w2 σ22
3.50
  −3/2 √  −3/2
π 1 2 π 1 1
,
2 σ12 w2 σ22 4 σ12 w2 σ22

for the first, second, third, and fourth integration, respectively.

Combining all of the obtained results for the integrals in 3.48 we achieve the result
stat-
ed as in the theorem.

Theorem 3.8. For nonnegative integer a and −1 < ρ < 1 the ath moment for W  X/Y is

 a 1  
1 σ1 a 3  a 
√  
EW a   √ ×Γ Γ − × 1 ρ 1 − 2 2 1 2a 2 . 3.51
σ2 2 σ2 2 2
14 Journal of Applied Mathematics

Proof. According to definition of expectation we can obtain the following result for the ath
moment:
∞  

π √  2 2 −3/2
EW a   wa 2 × 1 ρ 1 2 w σ2 σ12
0 2
 3.52
−3/2  −3/2 
−2ρ 2w2 σ22 σ12 w2 σ22 2σ12 dw.

√ √
By making transformation w  σ1 /σ2  m, w  σ1 /σ2  m/2 and w  σ1 /σ2  2m for the
first, second, and third parts of the component containing the integral we have

√  
σ1a 1

π  √  a 3 a
EW   3/2 × a 2 1 ρ 1 2 × B
a
,−
2 σ2 2 2
√  
ρ π σ a 1
 √  a 3 a
− a 4/2 × 1a 2 1 ρ 1 2 × B ,− 3.53
2 σ2 2 2
√  
ρ π σ1a 1
 √  a 3 a
− × 1 ρ 1 2 × B , − .
2−a 2/2 σ2a 2 2 2

Simplifying this we have the stated result of the theorem.

Theorem 3.9. The moment generating function of W is


∞ a  
t 1 σ1 a 1
MW t  × √
ao a! σ2 2 σ2
3.54
 
a 3  a 
√  
×Γ Γ − × 1 ρ 1 − 2 2 1 2a 2 .
2 2

Proof. The moment generating function of V at t is given by

 
MW t  E etw


∞ a
t
 EW a 
a0
a! 3.55
 a 1  

∞ a
t 1 σ1 a 3  a 
√  
 × √ ×Γ Γ − × 1 ρ 1 − 2 2 1 2a 2 .
ao a! σ2 2 σ2 2 2

3.3. Parameter Estimation of Rayleigh Mixture Distribution

We know the well-known method of the maximum likelihood estimation is very complicated
for the parameter estimation of mixture distribution and method of moment is very suitable
Journal of Applied Mathematics 15

in these cases. Hence, we used method of moments MoMs estimation of technique for es-
timation of the parameter of the Rayleigh mixture distribution.

3.3.1. Parameter Estimation of Rayleigh Mixture of Chi-Square Distribution

Let X1 , X2 , . . . , Xn be a random sample from the distribution defined in 3.2 where the para-
meter σ 2 is unknown.
The first sample raw moment is

1 n
 
m1  Xi  X say . 3.56
n i1

And as we already got

√
μ1  n σ 2π. 3.57

Hence
√
n σ 2π  X. 3.58

Therefore,

 2
X−n
σ! 2  . 3.59
2π

3.3.2. Parameter Estimation of Rayleigh Mixture of t-Distribution

Let X1 , X2 , . . . , Xn be a random sample from the distribution defined in 3.3 where the
para-
meter σ 2 is unknown. We want to estimate this parameter by method of moment.
The second sample raw moment is obtained as

1 n
 
m2  Xi2  S2 say . 3.60
n i1

We have already found

n  √
μ2  μ2  1 σ 2π , n > 2. 3.61
n−2

Hence, by the method of moments, we get

n  √
1 σ 2π  S2 , n > 2. 3.62
n−2
16 Journal of Applied Mathematics

Therefore,

 2
1 2 n − 2
σ! 
2
S −1 . 3.63
2π n

3.3.3. Parameter Estimation of Mixture of F-Distribution

Let X1 , X2 , . . . , Xn be a random sample from the distribution as specified in 3.4 where the
parameter σ 2 is unknown. We want to estimate this parameter by method of moments.
The first sample raw moment is

1 n
 
m1  Xi  X say . 3.64
n i1

And we already get

n2  √
μ1  n1 σ 2π . 3.65
n1 n2 − 2

Hence, from the method of moments estimator is

n2  √
n1 σ 2π  X. 3.66
n1 n2 − 2

Therefore,

  2  2
1 n1
σ! 2  Xn2 − 2 − n1 . 3.67
2π n2

4. Concluding Remarks
In this paper, we have presented the Rayleigh mixtures of distributions in which the weight
functions are assumed to be chi-square, t- and F-distributions, and the mixture of two cor-
related Rayleigh distributions has been presented. The moments, characteristic function and
shape characteristics of these mixtures distributions have also been studied. The Rayleigh
distribution is frequently used to model wave heights in oceanography and in communica-
tion theory to describe hourly median and instantaneous peak power of received radio sig-
nals. It could also be used to model the frequency of different wind speeds over a year at
wind turbine sites. The Rayleigh mixture of sampling distribution may be used in the similar
nature but with some additional informative environment. Suppose we want to know the dis-
tribution of the average fish caught by fisherman in the Bay of Bengal of a particular day.
Fishing depends on height of the wave and wind speed in that zone. As we know the average
amount fish catch by the fisherman depends on the weather of the Sea. If the wave heights are
very high the fishermen are prohibited to go to the sea for fishing if it is not so much danger-
ous but still the sea is unstable they are asked to be very careful during fishing. This means
Journal of Applied Mathematics 17

that average amount of fishing and standard deviation of the amount fish catch by the fisher-
men varies based on heights of the wave. The distribution of wave heights follows Rayleigh
distribution and distribution of average catch fish at a normal situation follows t-distri-
bution but at the Bay of Bangle it is seriously affected by height of wave; hence the average
number of fish catch at the Bay of Bangle will follow Rayleigh mixture of t-distribution.
Similarly, the distribution of the variability of the number of fishes catch by the fishermen at
the Bay of Bangle follows Rayleigh mixture of chi-square distribution. We hope the findings of
the paper will be useful for the practitioners that have been mentioned above.

Acknowledgment
The authors would like to thank the editor and referee for their useful comments and sugges-
tions which considerably improved the quality of the paper.

References
1 K. Pearson, “Contribution to the mathematical theory of evolution,” Philosophical Transactions of the
Royal Society A, vol. 185, pp. 71–110, 1894.
2 W. Mendenhall and R. J. Hader, “Estimation of parameters of mixed exponentially distributed failure
time distributions from censored life test data,” Biometrika, vol. 45, pp. 504–520, 1958.
3 V. Hasselblad, “Estimation of mixtures of distribution from the exponential family,” Journal of Ameri-
can Statistical Association, vol. 64, no. 328, pp. 300–314, 1968.
4 H. E. Daniels, “Mixtures of geometric distributions,” Journal of the Royal Statistical Society: Series B, vol.
23, pp. 409–413, 1961.
5 J. Chaine, “Une generalization dela loi binomiale negative,” Revue de Statistique Appliqué, vol. 13, no.
4, pp. 33–43, 1965.
6 M. K. Roy, S. Rahman, and M. M. Ali, “A class of poisson mixtured distributions,” Journal of
Information & Optimization Sciences, vol. 13, no. 2, pp. 207–218, 1992.
7 M. K. Roy, A. K. Sen, and M. M. Ali, “Binomial mixture of some distributions,” Journal of Information
& Optimization Sciences, vol. 13, no. 2, pp. 207–218, 1993.
8 M. K. Roy and S. K. Sinha, “Negative binomial mixture of normal moment distributions,” Metron, vol.
53, no. 3-4, pp. 83–91, 1995.
9 M. K. Roy and S. K. Sinha, “Negative binomial mixture of chi-square and F- distributions,” Journal of
Bangladesh Academy of Statistical Sciences, vol. 21, no. 1, pp. 25–34, 1997.
10 M. K. Roy, M. F. Imam, and J. C. Paul, “Gamma mixture of normal moment distribution,” International
Journal of Statistical Sciences, vol. 1, pp. 20–24, 2002.
11 M. K. Roy, M. R. Zaman, and N. Akter, “Chi-square mixture of gamma distribution,” Journal of Applied
Sciences, vol. 5, no. 9, pp. 1632–1635, 2005.
12 M. K. Roy, M. E. Haque, and B. C. Paul, “Earlang mixture of normal moment distribution,” Interna-
tional Journal of Statistical Sciences, vol. 6, pp. 29–37, 2007.
Copyright of Journal of Applied Mathematics is the property of Hindawi Publishing Corporation and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.