American Journal of Mathematics and Statistics 2017, 7(2): 51-59

DOI: 10.5923/j.ajms.20170702.01

Akshaya Distribution and Its Application

Rama Shanker

Department of Statistics, Eritrea Institute of Technology, Asmara, Eritrea

Abstract A new one parameter continuous distribution named “Akshaya distribution” for modeling lifetime data from
biomedical science and engineering has been proposed. Its mathematical and statistical properties including its shape,
moments, hazard rate function, mean residual life function, stochastic ordering, mean deviations, and Bonferroni and Lorenz
curves, have been presented. The conditions under which Akshaya distribution is over-dispersed, equi-dispersed and
under-dispersed are discussed along with some other one parameter lifetime distributions. The maximum likelihood
estimation and the method of moments have been discussed for estimating the parameter of the proposed distribution. Finally,
a numerical example of real lifetime data has been presented to test the goodness of fit of the proposed distribution and the fit
has been compared with other one parameter lifetime distributions.
Keywords Lifetime distribution, Moments, Index of dispersion, Hazard rate function, Mean residual life function,
Stochastic ordering, Mean deviations, Bonferroni and Lorenz curves, Estimation of parameter, Goodness of fit

1. Introduction a new lifetime distribution which is better than Akash,

The statistical analysis and modeling of lifetime data are Shanker, Amarendra, Aradhana, Sujatha, Devya, Shambhu,
essential in almost all applied sciences including, biomedical Lindley and exponential distributions for modeling lifetime
science, engineering, finance, and insurance, amongst others. data by considering a four-component mixture of an
The classical lifetime distributions namely exponential and exponential (θ ) , gamma ( 2,θ ) , gamma ( 3,θ ) and
Lindley (1958) distributions are popular in statistics for
gamma ( 4,θ ) distributions with their mixing proportions
modeling lifetime data. But these two classical lifetime
distributions are not suitable from theoretical and applied θ3 3θ 2
point of view. Shanker et al (2015) have done a critical and , ,
3 2 3 2
comparative study regarding the modeling of lifetime data
θ + 3 θ + 6 θ + 6 θ + 3θ + 6 θ + 6
using both exponential and Lindley distributions and found 6θ 6
and , respectively.
3
that there are several lifetime data where these classical θ + 3θ + 6θ + 6 2 3 2
θ + 3θ + 6θ + 6
lifetime distributions are not suitable due to their shapes, The probability density function (p.d.f.) of a new one
hazard rate functions and mean residual life functions, parameter lifetime distribution can be introduced as
amongst others. Recently, a number of one parameter
θ4
lifetime distributions have been introduced by Shanker = f ( x;θ ) 3 2
(1 + x )3 e−θ x ; x > 0, θ > 0 (1.1)
(2015 a, 2015 b, 2016 a, 2016 b, 2016 c, 2016 d, 2016 e) θ + 3θ + 6θ + 6
namely Akash, Shanker, Amarendra, Aradhana, Sujatha, We would call this distribution, “Akshaya distribution”.
Devya and Shambhu. Although these lifetime distributions The corresponding cumulative distribution function (c.d.f.)
give better fit than the classical exponential and Lindley of (1.1) can be obtained as
distributions, there are still some lifetime data where these
distributions are not suitable due to their theoretical and
F ( x; θ ) =1 − 1 + ( )
 θ 3 x3 + 3θ 2 (θ + 1) x 2 + 3θ θ 2 + 2θ + 2 x 
 e−θ x ; (1.2)
applied point of view.  θ 3 + 3θ 2 + 6θ + 6 
In search for a new lifetime distribution, we have proposed  
x > 0,θ > 0
* Corresponding author:
shankerrama2009@gmail.com (Rama Shanker) The graph of the p.d.f. and the c.d.f. of Akshaya
Published online at http://journal.sapub.org/ajms distribution for different values of the parameter θ are
Copyright © 2017 Scientific & Academic Publishing. All Rights Reserved shown in figures 1 and 2.
52 Rama Shanker: Akshaya Distribution and Its Application

pdf of Akshaya Distribution pdf of Akshaya Distribution

θ=0.2 θ=0.5 θ=0.8 θ=1
θ=1.5 θ=2 θ=2.5 θ=3
0.25 1
0.2 0.8
0.15 0.6

f(x)
f(x)

0.1 0.4
0.05 0.2
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
x x

Figure 1. Graph of the pdf of Akshaya distribution for different values of the parameter θ

CDF of Akshaya Distribution CDF of Akshaya Distribution

θ=0.2 θ=0.5 θ=0.8 θ=1 θ=1.5 θ=2 θ=2.5 θ=3
1 1
0.8 0.8
0.6 0.6
F(x)
F(x)

0.4 0.4
0.2 0.2
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
x x

Figure 2. Graph of the cdf of Akshaya distribution for different values of the parameter θ

2. Moments and Related Measures

The moment generating function of Akshaya distribution (1.1) can be obtained as
∞
θ4
M X (t ) 3 2 ∫
θ + 3θ + 6θ + 6 0
e ( ) 1 + 3 x + 3 x 2 + x3 dx
− θ −t x
( )
θ4  1 3 6 6 
=  + + + 
θ 3 + 3θ 2 + 6θ + 6  (θ − t ) (θ − t )2 (θ − t )3 (θ − t )4 

1 ∞  t  k k k k
θ4 3 ∞  k + 1  t  6 ∞  k + 2 t  6 ∞  k + 3 t 
=
θ 3 + 3θ 2 +=
 ∑   + ∑  
 θ  θ 2 k 0  k =
6θ + 6  θ k 0=
  + ∑  
 θ  θ 3 k 0  k =
  + ∑   
 θ  θ 4 k 0 k  θ 


∞ θ 3 + 3 ( k + 1)θ 2 + 3 ( k + 1)( k + 2 )θ + ( k + 1)( k + 2 )( k + 3)  t k
= ∑ θ 3 + 3θ 2 + 6θ + 6
 
θ 
k =0

tr
Thus the r th moment about origin, as given by the coefficient of in M X ( t ) , of Akshaya distributon (1.1) has been
r!
obtained as
r !θ 3 + 3 ( r + 1)θ 2 + 3 ( r + 1)( r + 2 )θ + ( r + 1)( r + 2 )( r + 3) 
µr′ =   ; r 1, 2,3,...
r 3 2
θ θ + 3θ + 6θ + 6 ( )
The first four moments about origin of Akshaya distribution (1.1) are obtained as

′
µ1 =
θ 3 + 6θ 2 + 18θ + 24 ,
′
µ2 =
2 θ 3 + 9θ 2 + 36θ + 60
, ( )
(
θ θ 3 + 3θ 2 + 6θ + 6 )
θ 2 θ 3 + 3θ 2 + 6θ + 6 ( )
 American Journal of Mathematics and Statistics 2017, 7(2): 51-59 53

µ3′ =
(
6 θ 3 + 12θ 2 + 60θ + 120 ), µ4′ =
(
24 θ 3 + 15θ 2 + 90θ + 210 )
3
( 3
θ θ + 3θ + 6θ + 6 2
) θ 4
(θ 3 2
+ 3θ + 6θ + 6 )
Thus the moments about mean of the Akshaya distribution (1.1) are obtained as
θ 6 + 12θ 5 + 66θ 4 + 192θ 3 + 288θ 2 + 288θ + 144
µ2 =
( )
2
θ 2 θ 3 + 3θ 2 + 6θ + 6

µ3 =
(
2 θ 9 + 18θ 8 + 144θ 7 + 630θ 6 + 1620θ 5 + 2916θ 4 + 3888θ 3 + 3888θ 2 + 2592θ + 864 )
( )
3
θ 3 θ 3 + 3θ 2 + 6θ + 6
 θ 12 + 24θ 11 + 260θ 10 + 1632θ 9 + 6648θ 8 + 19680θ 7 + 44496θ 6 + 78336θ 5 
9 
 +106992θ 4 + 110592θ 3 + 82944θ 2 + 41472θ + 10368 
µ4 =  
( )
4 3 2 4
θ θ + 3θ + 6θ + 6

The coefficient of variation ( C.V ) , coefficient of skewness ( )

β1 , coefficient of kurtosis ( β2 ) and index of
dispersion ( γ ) of Akshaya distribution (1.1) are thus obtained as

σ θ 6 + 12θ 5 + 66θ 4 + 192θ 3 + 288θ 2 + 288θ + 144

C=
.V =
µ1′ θ 3 + 6θ 2 + 18θ + 24
 θ 9 + 18θ 8 + 144θ 7 + 630θ 6 + 1620θ 5 + 2916θ 4 + 3888θ 3 
2 
µ3  +3888θ 2 + 2592θ + 864 
β1 =
=  
µ23/2
( )
6 5 4 3 2 3/2
θ + 12θ + 66θ + 192θ + 288θ + 288θ + 144
 θ 12 + 24θ 11 + 260θ 10 + 1632θ 9 + 6648θ 8 + 19680θ 7 + 44496θ 6 + 78336θ 5 
9 
µ4  +106992θ 4 + 110592θ 3 + 82944θ 2 + 41472θ + 10368 
β2 =
=  
µ22
( )
6 5 4 3 2 2
θ + 12θ + 66θ + 192θ + 288θ + 288θ + 144

σ2 θ 6 + 12θ 5 + 66θ 4 + 192θ 3 + 288θ 2 + 288θ + 144

γ
= =
µ1′ ( )(
θ θ 3 + 3θ 2 + 6θ + 6 θ 3 + 6θ 2 + 18θ + 24 )
Table 1. Over-dispersion, equi-dispersion and under-dispersion of Akshaya, Akash, Shanker, Amarendra, Aradhana, Sujatha, Devya, Shambhu, Lindley
and exponential distributions for parameter θ

Distribution Over-dispersion (µ < σ ) 2

Equi-dispersion (µ = σ ) 2
Under-dispersion (µ > σ ) 2

Akshaya θ < 1.327527885 θ = 1.327527885 θ > 1.327527885

Akash θ < 1.515400063 θ = 1.515400063 θ > 1.515400063
Shanker θ < 1.171535555 θ = 1.171535555 θ > 1.171535555
Amarendra θ < 1.525763580 θ = 1.525763580 θ > 1.525763580
Aradhana θ < 1.283826505 θ = 1.283826505 θ > 1.283826505
Sujatha θ < 1.364271174 θ = 1.364271174 θ > 1.364271174
Devya θ < 1.451669994 θ = 1.451669994 θ > 1.451669994
Shambhu θ < 1.149049973 θ = 1.149049973 θ > 1.149049973
Lindley θ < 1.170086487 θ = 1.170086487 θ > 1.170086487
Exponential θ <1 θ =1 θ >1
54 Rama Shanker: Akshaya Distribution and Its Application

The condition under which Akshaya distribution is over-dispersed, equi-dispersed, and under-dispersed has been studied
along with condition under which Akash, Shanker, Amarendra, Aradhana, Sujatha, Devya, Shambhu, Lindley and
exponential distributions are over-dispersed, equi-dispersed, and under-dispersed and are presented in table 1.

3. Hazard Rate Function and Mean Residual Life Function

Let X be a continuous random variable with p.d.f. f ( x ) and c.d.f. F ( x ) . The hazard rate function (also known as
the failure rate function) and the mean residual life function of X are respectively defined as
P ( X < x + ∆x X > x ) f ( x)
=h( x) lim
= (3.1)
∆x →0 ∆x 1 − F ( x)

1 ∞
and m (=
x ) E  X − x X >=
x  ∫ 1 − F ( t )  dt (3.2)
1 − F ( x) x 

The corresponding hazard rate function, h ( x ) and the mean residual life function, m ( x ) of the Akshaya distribution
(1.1) are obtained as
3
θ 4 (1 + x ) (3.3)
h( x) =
(
θ 3 x3 + 3θ 2 (θ + 1) x 2 + 3θ θ 2 + 2θ + 2 x + θ 3 + 3θ 2 + 6θ + 6 ) ( )
and
1
m( x) =
θ x + 3θ

3 3 2
(θ + 1) x 2
( 2
) (
+ 3θ θ + 2θ + 2 x + θ 3 + 3θ 2 + 6θ + 6  e−θ x
 )
∞
(
× ∫ θ 3t 3 + 3θ 2 (θ + 1) t 2 + 3θ θ 2 + 2θ + 2 t + θ 3 + 3θ 2 + 6θ + 6  e −θ t dt
  ) ( )
x

=
( ) (
θ 3 x3 + 3x 2 + 3x + 1 + 6θ 2 x 2 + 2 x + 1 + 18θ ( x + 1) + 24 ) (3.4)

3 3 2 2
(
θ θ x + 3θ (θ + 1) x + 3θ θ + 2θ + 2 x + θ 3 + 3θ 2 + 6θ + 6 2
) ( )
4 3
seen that h ( 0 )
θ
and m ( 0 ) θ= + 6θ 2 + 18θ + 24
It can be easily= = =f ( 0 ) µ1′ . It is also obvious
θ 3 + 3θ 2 + 6θ + 6 θ θ 3 + 3θ 2 + 6θ + 6 ( )
from the graphs of h ( x ) and m ( x ) that h ( x ) is an increasing function of x , and θ , whereas m ( x ) is firstly
increasing and then decreasing function of x , and θ .
The graphs of the hazard rate function and mean residual life function of Akshaya distribution (1.1) are shown in figures 3
and 4.

h(x) of Akshaya Distribution h(x) of Akshaya Distribution

θ=0.2 θ=0.5 θ=0.8 θ=1 θ=1.5 θ=2 θ=2.5 θ=3
1 3
0.8 2.5
0.6 2
h(x)

h(x)

1.5
0.4
1
0.2 0.5
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
x x

Figure 3. Graph of hazard rate function of Akshaya distribution for varying values of parameter θ
 American Journal of Mathematics and Statistics 2017, 7(2): 51-59 55

m(x) of Akshaya Distribution m(x) of Akshaya Distribution

θ=0.2 θ=0.5 θ=0.8 θ=1 θ=1.5 θ=2 θ=2.5 θ=3
100 6
80 5
4
60

m(x)
m(x)

3
40
2
20 1
0 0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
x x

Figure 4. Graph of mean residual life function of Akshaya distribution for varying values of parameter θ

4. Stochastic Orderings
Stochastic ordering of positive continuous random variable is an important tool for judging their comparative behavior. A
random variable X is said to be smaller than a random variable Y in the
(i) stochastic order ( X ≤ st Y ) if FX ( x ) ≥ FY ( x ) for all x
(ii) hazard rate order ( X ≤ hr Y ) if hX ( x ) ≥ hY ( x ) for all x

(iii) mean residual life order ( X ≤mrl Y ) if m X ( x ) ≤ mY ( x ) for all x

f X ( x)
(iv) likelihood ratio order ( X ≤lr Y ) if decreases in x .
fY ( x )
The following results due to Shaked and Shanthikumar (1994) are well known for establishing stochastic ordering of
distributions
X ≤lr Y ⇒ X ≤ hr Y ⇒ X ≤ mrl Y
⇓
X ≤ st Y

The Akshaya distribution is ordered with respect to the strongest ‘likelihood ratio’ ordering as shown in the following
theorem:
Theorem: Let X ∼ Akshaya distribution (θ1 ) and Y ∼ Akshaya distribution (θ 2 ) . If θ1 > θ 2 , then X ≤lr Y and
hence X ≤ hr Y , X ≤ mrl Y and X ≤ st Y .
Proof: We have

f X ( x)
=
(
θ14 θ 23 + 3θ 22 + 6θ 2 + 6) e−(θ −θ )x 1 2 ; x>0
fY ( x )
θ 24 (θ13 + 3θ12 + 6θ1 + 6 )

Now

=
f ( x)
log fX x log 
2(
 θ 4 θ 3 + 3θ 2 + 6θ + 6
1 2 2 )  − (θ − θ ) x .
Y( )
 2 1 (
 θ θ + 3θ + 6θ + 6
4 3
1
2
1 )  1 2
d f X ( x)
This gives log fY ( x )
− (θ1 − θ 2 ) .
=
dx
d f X ( x)
Thus for θ1 > θ 2 , log fY ( x )
≤ 0 . This means that X ≤lr Y and hence X ≤ hr Y , X ≤ mrl Y and X ≤ st Y .
dx
56 Rama Shanker: Akshaya Distribution and Its Application

5. Mean Deviations about Mean and Median

The amount of scatter in a population is measured to some extent by the totality of deviations usually from their mean and
median. These are known as the mean deviation about the mean and the mean deviation about the median and are defined as
∞ ∞
δ1 ( X=) ∫ x − µ f ( x ) dx and δ 2 ( X=) ∫ x − M f ( x ) dx , respectively, where µ = E ( X ) and M = Median ( X ) .
0 0
The measures δ1 ( X ) and δ 2 ( X ) can be computed using the following simplified relationships
µ ∞
δ1 ( X ) = ∫ ( µ − x ) f ( x ) dx + ∫ ( x − µ ) f ( x ) dx
0 µ
µ ∞
= µ F ( µ ) − ∫ x f ( x ) dx − µ 1 − F ( µ )  + ∫ x f ( x ) dx
0 µ
∞
= 2 µ F ( µ ) − 2 µ + 2 ∫ x f ( x ) dx
µ
µ
= 2 µ F ( µ ) − 2 ∫ x f ( x ) dx (5.1)
0

and
M ∞
δ2 ( X ) = ∫ ( M − x ) f ( x ) dx + ∫ ( x − M ) f ( x ) dx
0 M
M ∞
= M F (M ) − ∫ x f ( x ) dx − M 1 − F ( M ) + ∫ x f ( x ) dx
0 M
∞
=− µ + 2 ∫ x f ( x ) dx
M
M
= µ − 2 ∫ x f ( x ) dx (5.2)
0
Using p.d.f. (1.1) and expression for the mean of Akshaya distribution, we get

( )
θ 4 µ 4 + 3µ 3 + 3µ 2 + µ + θ 3 4 µ 3 + 9 µ 2 + 6 µ + 1
 ( ) e−θ µ
 
µ
 (
+6θ 2 2 µ 2 + 3µ + 1 + 6θ ( 4 µ + 3) + 24) 

∫ x f ( x ) dx= µ − (5.3)
0 θ θ 3 + 3θ 2 + 6θ + 6 ( )
(
θ 4 M 4 + 3M 3 + 3M 2 + M + θ 3 4 M 3 + 9 M 2 + 6 M + 1 
 )
 −θ M ( )
 e
M

2
(
+6θ 2 M + 3M + 1 + 6θ ( 4 M + 3) + 24
2
) 

∫ x f ( x ) dx
= µ − (5.4)
0
3 2
θ θ + 3θ + 6θ + 6 ( )
Using expressions (5.1), (5.2), (5.3) and (5.4), the mean deviation about mean δ1 ( X ) and the mean deviation about
median δ 2 ( X ) of Akshaya distribution (1.1), after a little algebraic simplification, are obtained as
 American Journal of Mathematics and Statistics 2017, 7(2): 51-59 57

{ ( ) (
 θ 3 µ 3 + 3µ 2 + 3µ + 1 + 6θ 2 µ 2 + 2 µ + 1 + 18θ ( µ + 1) + 24
δ1 ( X ) = 2 
) }  e −θ µ (5.5)
 (
θ θ 3 + 3θ 2 + 6θ + 6 ) 


( )
θ 4 M 4 + 3M 3 + 3M 2 + M + θ 3 4 M 3 + 9 M 2 + 6 M + 1 
  −θ M ( )
2 e

δ2 ( X )  ( )
+6θ 2 M + 3M + 1 + 6θ ( 4 M + 3) + 24
2 2

 −µ (5.6)
3 2
θ θ + 3θ + 6θ + 6 ( )

6. Bonferroni and Lorenz Curves

The Bonferroni and Lorenz curves (Bonferroni, 1930) and Bonferroni and Gini indices have applications not only in
economics to study income and poverty, but also in other fields like reliability, demography, insurance and medicine. The
Bonferroni and Lorenz curves are defined as

1   1  
q ∞ ∞ ∞
1
B ( p ) =∫ x f ( x ) dx = ∫ x f ( x ) dx − ∫ x f ( x ) dx  = µ − ∫ x f ( x ) dx  (6.1)
pµ 0 pµ  0  pµ  
 q   q 

1  1 
q ∞ ∞ ∞
1
and L ( p ) = ( ) ( ) ( ) ( )
µ ∫0 µ  ∫0 ∫ ∫
x f x dx =  x f x dx − x f x dx  = µ − x f x dx  (6.2)
 µ 
 q   q 
respectively or equivalently
p
1
B ( p) = F −1 ( x ) dx (6.3)
p µ ∫0
p
and L ( p ) = 1 F −1 ( x ) dx (6.4)
µ ∫0

respectively, where µ = E ( X ) and q = F −1 ( p ) .

The Bonferroni and Gini indices are thus defined as
1
B = 1 − ∫ B ( p ) dp (6.5)
0
1
and G = 1 − 2 L ( p ) dp (6.6)
∫
0

respectively.
Using p.d.f. of Akshaya distribution (1.1), we get

(
θ 4 q 4 + 3q3 + 3q 2 + q + θ 3 4q3 + 9q 2 + 6q + 1
 ) ( ) e−θ q
 
∞
 ( )
+6θ 2 2q 2 + 3q + 1 + 6θ ( 4q + 3) + 24 

(6.7)
∫ x f ( x ) dx =
q (
θ θ 3 + 3θ 2 + 6θ + 6 )
Now using equation (6.7) in (6.1) and (6.2), we get

(
 θ 4 q 4 + 3q3 + 3q 2 + q + θ 3 4q3 + 9q 2 + 6q + 1
  ) ( ) e−θ q 
  

B (=
p)
2
(
1  +6θ 2q + 3q + 1 + 6θ ( 4q + 3) + 24
2
) 



1 −  (6.8)
p θ 3 + 6θ 2 + 18θ + 24 
 
 
 
 58 Rama Shanker: Akshaya Distribution and Its Application

( ) (
θ 4 q 4 + 3q3 + 3q 2 + q + θ 3 4q3 + 9q 2 + 6q + 1
 ) e−θ q
 
( )
+6θ 2 2q 2 + 3q + 1 + 6θ ( 4q + 3) + 24
L ( p )= 1 − 


(6.9)

θ 3 + 6θ 2 + 18θ + 24
Now using equations (6.8) and (6.9) in (6.5) and (6.6), the Bonferroni and Gini indices of Akshaya distribution (1.1) are
obtained as

( ) (
θ 4 q 4 + 3q3 + 3q 2 + q + θ 3 4q3 + 9q 2 + 6q + 1
 ) e−θ q
 
( )
+6θ 2 2q 2 + 3q + 1 + 6θ ( 4q + 3) + 24
B= 1− 


(6.10)

θ 3 + 6θ 2 + 18θ + 24

( ) (
θ 4 q 4 + 3q3 + 3q 2 + q + θ 3 4q3 + 9q 2 + 6q + 1

2
) e−θ q


G  ( )
+6θ 2 2q 2 + 3q + 1 + 6θ ( 4q + 3) + 24 
 −1
(6.11)

θ 3 + 6θ 2 + 18θ + 24

7. Estimation of Parameter lifetime data set from biomedical science and engineering. In
this section, we present the goodness of fit of Akshaya
7.1. Maximum Likelihood Estimate (MLE) distribution to a real lifetime data and compare its goodness
of fit with the one parameter Akash, Shanker, Amarendra,
Let ( x1 , x2 , x3 , ... , xn ) be a random sample from
Aradhana, Sujatha, Devya, Shambhu, Lindley and
Akshaya distribution (1.1). The likelihood function, L of exponential distributions.
Akshaya distribution is given by The data set represents the lifetime’s data relating to relief
n times (in minutes) of 20 patients receiving an analgesic and
 θ4  n
3 − nθ x
L  3
 θ + 3θ 2 + 6θ + 6  ∏ (1 + xi ) e reported by Gross and Clark (1975, P. 105). The data are as
follows:
  i =1

The natural log likelihood function is thus obtained as 1.1, 1.4, 1.3, 1.7, 1.9, 1.8, 1.6, 2.2, 1.7, 2.7, 4.1, 1.8,
1.5, 1.2, 1.4, 3.0, 1.7, 2.3, 1.6, 2.0
 θ4  n

 θ + 3θ 2 + 6θ + 6  ∑ (
=ln L n ln  3  + 3 ln 1 + xi ) − n θ x In order to compare lifetime distributions, −2ln L , AIC
  i =1 (Akaike Information Criterion) and K-S Statistics
(Kolmogorov-Smirnov Statistics) for the above data set have
Now d ln L = −
2
4n 3n θ + 2θ + 2 ( − nx
) been computed. The formulae for computing AIC and K-S
dθ θ θ 3 + 3θ 2 + 6θ + 6 Statistics are as follows:
where x is the sample mean. AIC = , K -S Sup Fn ( x ) − F0 ( x ) ,
−2ln L + 2k=
x
The maximum likelihood estimate, θˆ of θ is the
where k = the number of parameters, n = the sample size
d ln L
solution of the equation = 0 and so it can be obtained and Fn ( x ) is the empirical distribution function.
dθ
by solving the following fourth degree polynomial equation
The best distribution is the distribution which corresponds
xθ 4 + ( 3x − 1)θ 3 + 6 ( x − 1)θ 2 + 6 ( x − 3)θ − 24 =
0 (7.1.1) to lower values of −2ln L , AIC, and K-S statistics. The

7.2. Method of Moment Estimate (MOME)

MLE (θˆ ) and standard error, S.E θˆ () of θ , −2ln L ,

Equating the population mean of the Akshaya distribution
to the corresponding sample mean, the MOME θ of θ is
same as given by equation (7.1.1)
8. Applications and Goodness of Fit
The Akshaya distribution has been fitted to a number of
AIC and K-S Statistic of the fitted distributions are presented
in the following table 2.
It can be easily seen from above table that Akshaya
distribution gives better fit than the fit given by Akash,
Shanker, Amarendra, Aradhana, Sujatha, Devya, Shambhu,
Lindley and exponential distributions and hence it can be
considered as an important lifetime distribution over these
lifetime distributions.
 American Journal of Mathematics and Statistics 2017, 7(2): 51-59 59

Table 2. MLE’s, S.E (θˆ ) - 2ln L, AIC and K-S Statistics of the fitted
distributions of the given data

MLE S.E
K-S
REFERENCES
Distributions
(θˆ ) ( )
θˆ −2 ln L AIC
statistic [1] Bonferroni, C.E. (1930): Elementi di Statistca generale,
Seeber, Firenze.
Akshaya 1.441686 0.167924 53.01 55.01 0.430
Akash 1.156920 0.145551 59.52 61.52 0.442 [2] Gross, A.J. and Clark, V.A. (1975): Survival Distributions:
Reliability Applications in the Biometrical Sciences, John
Shanker 0.803876 0.119299 59.78 61.78 0.442
Wiley, New York.
Amarendra 1.480769 0.120356 55.64 57.64 0.465
[3] Lindley, D.V. (1958): Fiducial distributions and Bayes’
Aradhana 1.123193 0.152614 56.37 58.37 0.453
theorem, Journal of the Royal Statistical Society, Series B, 20,
Sujatha 1.136745 0.149841 57.49 59.49 0.442 102- 107.
Devya 1.841919 0.169223 54.50 56.50 0.548
[4] Shaked, M. and Shanthikumar, J.G. (1994): Stochastic Orders
Shambhu 2.215392 0.176084 53.89 55.89 0.504 and Their Applications, Academic Press, New York.
Lindley 0.816112 0.136090 60.49 62.49 0.460
[5] Shanker, R., Hagos, F, and Sujatha, S. (2015): On modeling
Exponential 0.526314 0.117687 65.67 67.67 0.471 of Lifetimes data using exponential and Lindley distributions,
Biometrics & Biostatistics International Journal, 2 (5), 1-9.
[6] Shanker, R. (2015 a): Akash distribution and its Applications,
9. Conclusions International Journal of Probability and Statistics, 4(3), 65 –
75.
A new one parameter lifetime distribution named,
“Akshaya distribution” has been suggested for modeling [7] Shanker, R. (2015 b): Shanker distribution and its
lifetime data from engineering and medical science. Its Applications, International Journal of Statistics and
Applications, 5(6), 338 – 348.
important mathematical and statistical properties including
shape, moments, skewness, kurtosis, hazard rate function, [8] Shanker, R. (2016 a): Amarendra distribution and its
mean residual life function, stochastic ordering, mean Applications, American Journal of Mathematics and
deviations, Bonferroni and Lorenz curves have been Statistics, 6(1), 44 – 56.
discussed. The method of moments and the method of [9] Shanker, R. (2016 b): Aradhana distribution and its
maximum likelihood estimation have also been discussed for Applications, International Journal of Statistics and
estimating its parameter. Finally, the goodness of fit using Applications, 6(1), 23 – 34.
K-S Statistics (Kolmogorov-Smirnov Statistics) for a real [10] Shanker, R. (2016 c): Sujatha distribution and its
lifetime data have been presented to demonstrate its Applications, Statistics in Transition-new series, 17(3), 1 –
applicability over Akash, Shanker, Amarendra, Aradhana, 20.
Sujatha, Devya, Shambhu, Lindley and exponential
[11] Shanker, R. (2016 d): Devya distribution and its Applications,
distributions. International Journal of Statistics and Applications, 6(4), 189
– 202.

ACKNOWLEDGEMENTS [12] Shanker, R. (2016 e): Shambhu distribution and its

Applications, International Journal of Probability and
The author is grateful to the editor- in - chief and the Statistics, 5(2), 48 -63.
anonymous reviewer for helpful comments which improved
the quality of the paper.