A NEW EXTENSION OF DEVYA DISTRIBUTION WITH
PROPERTIES AND APPLICATIONS IN REAL LIFE DATA
M. Sajish Kumar1 and C. Subramanian2
1,2Department of Statistics, Annamalai University, Annamalai nagar, Tamil Nadu, India
1Corresponding Author: [email protected]
Abstract
In this paper, we proposed a new probability model called as the Length Biased Devya
distribution. The statistical properties of this distribution including, the mean, variance,
harmonic mean and probability generating functions have been studied. The maximum
likelihood method has been used to estimate the parameter. Finally a real life data has been
analysed for illustration purpose.
Keywords: Devya distribution, weighted distribution, Reliability Analysis, Order
Statistics, Maximum Likelihood Estimation.
1. INTRODUCTION
Weighted distribution provides an approach to dealing with model specification and data
interpretation problems. Fisher (1934) and Rao (1965) introduced and unified the concept of
weighted distribution. When the probability of observing a positive-valued random variable is
proportional to the value of the variable the resultant is length-biased distribution. A table for
some basic distributions and their length-biased forms is given by Patil and Rao (1978) such
as Lognormal, Gamma, Pareto, Beta distribution. Modi and Gill (2015), obtained length
biased weighted Maxwell distribution and its applications. Shenbagaraja, et al. (2019),
discussed on length biased Garima distribution with several properties and its applications.
Mudasir and Ahmad (2018), discussed on characterization and estimation of the length biased
Nakagami distribution. Recently, Rather and Subramanian (2019) discussed on length biased
erlang truncated exponential distribution with life time data which shows more flexibility
than classical distributions.
A new one parametric lifetime distribution named Devya distribution was given by
Shanker (2016). Its statistical and mathematical properties including moments, order
statistics, Bonferroni and Lorenz curves have been discussed. In this paper, we will discuss
the length biased version of Devya distribution with its properties and applications.
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1188
2. LENGTH BIASED DEVYA (LBD) DISTRIBUTION
The probability density function (pdf) of the Devya Distribution is given by
)1(0,0;1)2462(
),( 432
234
5
xexxxxxf x
and the cumulative density function (cdf) of the Devya Distribution is given by
)2(,0;
)2462(
24)2(6)234()(11),(
234
222332344
xe
xxxxxxxxxxxF
x
and its mean is
)3()2462(
1202462)(
234
234
XE
Suppose X is a non- negative random variable with probability density function )(xf .Let
)(xw be the nonnegative weight function, then the probability density function of the
weighted random variable Xw is given
)4(0;))((
)()()( x
xwE
xfxwxf w
Where w(x) be a non-negative weight function and )()()( xfxwXE .
For different weighted models, we have different choices of the weight function w(x).when cx the resulting distribution is termed as weighted distribution. In this paper, we have to find
the Length biased version of Devya distribution, so we will take c = 1 in weightscxxw )( ,
in order to get the Length biased Devya distribution and its probability density function (pdf)
is given by:
)5()(
)()(
xE
xxfxf l
Substitute, the values of equation (1) and (3) in equation (5), we will get the probability
density function of Length biased Devya distribution.
)6(0,0;1)1202462(
),( 432
234
6
xexxxxxxf x
l
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1189
Now, the cumulative distribution function of the Length biased Devya distribution is obtained
as
x
ll dxxfxF0
)(),(
x
x
l xdxexxxxxxF0
432
234
6
0,0;1)1202462(
),(
After simplification, we will get the cumulative distribution function of Length biased Devya
distribution as
)7(
;61
;51
;41
;31
;2
)1202462()(
43
2
234
4
xx
xxx
xFl
3. RELIABILITY ANALYSIS
In this section, we will discuss about the survival function, failure rate, reverse hazard rate
and Mills ratio of the length biased Devya distribution.
The survival function or the reliability function of the Length biased Devya
distribution is given by
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1190
)(1)( xFxS l
xx
xxx
xS
;61
;51
;41
;31
;2
)1202462(1)(
43
2
234
4
The hazard function is also known as the hazard rate, instantaneous failure rate or force of
mortality and is given by
)(1
)()(
xF
xfxh
l
l
xx
xxx
exxxxxxh
x
;61
;51
;41
;31
;2
)1202462(
1)(
43
2
4234
4324
and Reverse hazard rate and Mills Ratio of the Length biased Devya distribution is given by
)(
)()(
xF
xfxh
l
l
r
xxxxx
exxxxxxh
x
r
;61
;51
;41
;31
;2
1)(
432
4324
)(
1
xhMillsRatio
r
xexxxxx
xxxxx
)1(
;61
;51
;41
;31
;2
4324
432
3.1 Harmonic Mean
The Harmonic mean of the Length Biased Devya distribution model can be obtained as
xEMH
1.
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1191
dxxfx
l )(1
0
dxexxxxxx
x
0
432
234
6
1)1202462(
1
0
432
234
6
1)1202462(
dxexxxx x
0
4
0
3
0
2
00
234
6
)1202462(
dxexdxex
dxexdxxedxe
xx
xxx
on simplification, we get
)1202462(
)2462(..
234
234
MH
4. MOMENTS AND ASSOCIATED MEASURES
Let X denotes the random variable of Length biased Devya distribution then the thr order
moment )( rXE of Length biased Devya distribution is obtained as
dxexxxxxx xr
432
234
6
0
1)1202462(
)8(1202462
)!5()!4()!3()!2()!1(234
234
r
rrrrr
Putting r =1 in equation (8), we will get the mean of Length biased Devya distribution which
is given by
1202462
7201202462234
234
1
dxxfxXE l
r
r
r )()(0
'
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1192
and putting r = 2 in equation (8), we get the second moment of Length biased Devya
distribution as
1202462
50407201202462342
234
2
therefore,
2
12 )(Variance
2
234
234
2342
234
1202462
7201202462
1202462
5040720120246
2
234
234
2342
234
1202462
7201202462
1202462
5040720120246).(.
DS
Coefficient of Variation (C.V) ='
1
7201202462
720120246212024625040720120246234
2234234234
Coefficient of Dispersion (γ) = '
1
2
12024627201202462
720120246212024625040720120246234234
2234234234
5. MOMENT GENERATING FUNCTION AND CHARACTERISTIC
FUNCTION
Let X have a Length biased Devya distribution, then the MGF of X is obtained as
0
)()()( dxxfeeEtM w
txtx
X
Using Taylor’s expansion
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1193
0
2
)(!2
)(1)()( dxxf
txtxeEtM l
tx
X
0 0
)(!
dxxfxj
tl
j
j
j
'!0
j
j
j
j
t
)9(1202462
)!5()!4()!3()!2()!1(
! 234
234
0
j
j
j jjjjj
j
t
Similarly, the characteristic function of Length biased Devya distribution can be obtained as
)()( itMt XX
0234
234
)10(!1202462
)!5()!4()!3()!2()!1(
j
j
j j
itjjjjj
6. ORDER STATISTICS
Let X(1), X(2), ...,X(n) be the order statistics of a random sample X1, X2, ..., Xn drawn from the
continuous population with probability density function fx(x) and cumulative density function
with Fx(x), then the pdf of rthorder statistics X(r) is given by
rn
X
r
XXrX xFxFxfrnr
nxf
)(1)()(
)!()!1(
!)(
1
)( (11)
Using the equations (6) and (7) in equation (11), the probability density function of rth order
statistics X(r) of Length biased Devya distribution is given by
rn
r
x
rX
xxxxx
xxxxx
exxxxxnr
nxf
;61
;51
;41
;31
;2)1202462(
1
;61
;51
;41
;31
;2)1202462(
1)1202462()!1()!1(
!)(
432234
6
1
432234
6
432
234
6
)(
Therefore, the probability density function of higher order statistics X(n) can be obtained as
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1194
1
432234
6
432
234
6
)(
;61
;51
;41
;31
;2)1202462(
1)1202462(
)(
n
x
nX
xxxxx
exxxxxnxf
and the pdf of 1st order statistic X(1)can be obtained as
1
432234
6
432
234
6
)1(
;61
;51
;41
;31
;2)1202462(
1
1)1202462(
)(
n
x
X
xxxxx
exxxxxxf
7. ENTROPIES
The concept of entropy is important in different areas such as probability and statistics,
physics, communication theory and economics. Entropies quantify the diversity, uncertainty,
or randomness of a system. Entropy of a random variable X is a measure of variation of the
uncertainty.
7.1 Renyi Entropy
The Renyi entropy is important in ecology and statistics as index of diversity. The Renyi
entropy is also important in quantum information, where it can be used as a measure of
entanglement. For a given probability distribution, Renyi entropy is given by
dxxfe )(log
1
1)(
; where β > 0, and β ≠1
0
432
234
6
1)1202462(1
1)( dxexxxxxLoge x
iiiii
0
432
234
6
112024621
1)( dxexxxxxLoge x
iiiii
)12()1(
12024621
1)(
0
432
234
6
dxexxxxxLoge x
iiii
Using binomial expansion in equation (12), we get
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1195
dxex
l
k
k
j
j
i
ie xlkji
i j k l
0
11
0 0 0 0234
6
1202462log
1
1)(
10 0 0 0
234
6 1
1202462log
1
1)(
lkji
i j k l
lkji
l
k
k
j
j
i
ie
7.2 Tsallis Entropy
A generalization of Boltzmann-Gibbs (B-G) statistical mechanics initiated by Tsallis has
focussed a great deal to attention. This generalization of B-G statistics was proposed firstly
by introducing the mathematical expression of Tsallis entropy (Tsallis, 1988) for a
continuous random variable is defined as follows
0
)(11
1dxxfS
0
432
234
6
1)1202462(
11
1dxexxxxxS x
iiiii
0
432
234
6
11202462
11
1dxexxxxxS x
iiiii
)13()1(
12024621
1
1
0
432
234
6
dxexxxxxS x
iiii
Using binomial expansion to equation (13), we get
dxex
l
k
k
j
j
i
iS xlkji
i j k l
0
11
0 0 0 0234
6
12024621
1
1
10 0 0 0
234
6 1
12024621
1
1lkji
i j k l
klji
l
k
k
j
j
i
iS
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1196
8. BONFERRONI AND LORENZ CURVES
The Bonferroni and Lorenz curves (Bonferroni, 1930) 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
q
dxxxfp
pB01
)('
1)(
and
q
dxxxfpL01
)('
1)(
where 1202462
7201202462234
234/
1
and
)(1 pFq
dxexxxxxp
pB x
q
0
432
234
1
6
1)1202462(
)(
dxexdxexdxexdxexdxexp
pB x
q
x
q
x
q
x
q
x
q
0
5
0
4
0
3
0
2
0
234
1
6
)1202462()(
dxetdxet
dxetdxetdxet
ppB
x
q
x
q
x
q
x
q
x
q
0
5
6
0
4
5
0
3
4
0
2
3
0
2
234
1
6
11
111
)1202462()(
dxetdxet
dxetdxetdxet
ppB
x
q
x
q
x
q
x
q
x
q
0
16
6
0
15
5
0
14
4
0
13
3
0
12
2
234
1
6
11
111
)1202462()(
);!5(
1);!4(
1);!3(
1;!2(
1;1(
)7201202462()(
432234
5
qqqqqp
pB
L(p)=p(B(p)
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1197
);!5(
1);!4(
1);!3(
1;!2(
1;1(
)7201202462()(
432234
5
qqqqqpL
9. MAXIMUM LIKELIHOOD ESTIMATOR
Let ).........,,,( 321 nxxxx be a random sample of size from Devya distribution. The likelihood
function, of is given by
)14(1()1202462(
);(1
432
234
6
n
i
x
iiiii
n
iexxxxxxL
The natural log likelihood function thus obtained as
)15(1ln
)1202462(ln);(ln
0
432
234
6
i
i
iiiii xxxxxxnxL
The maximum likelihood estimates of θ can be obtained by differentiating equation (15) with
respect to θ and must satisfy the normal equation
XnxxxxxnL
iiiii
4322346 1lnln)1202462ln((lnlog
)16(0)1202462(
2412646log234
23
Xnn
L
Because of the complicated form of likelihood equations (16), algebraically it is very
difficult to solve the system of nonlinear equations. Therefore, we use R and wolfram
mathematica for estimating the required parameters.
10. LIKELIHOOD RATIO TEST
Let X1, X2, ... , Xn be a random sample from the LBD distribution. To test the hypothesis
);()(:against);()(: 1 xfxfHxfxfH lo
In order to test whether the random sample of size n comes from the Devya distribution or
LBD distribution, the following test statistic is used
n
i
l
xf
xf
L
L
10
1
);(
);(
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1198
i
n
i
x
1234
234
)1202462(
)2462((
n
i
i
n
x1
234
234
)1202462(
)2462(
We reject the null hypothesis if
n
i
i
n
kx1
234
234
)1202462(
)2462(
n
n
i
i kx
)2462(
)1202462(or,
234
234
1
*
n
n
i
i kkkx
)2462(
)1202462(where,
234
234**
1
*
For large sample size n, 2 log∆ is distributed as chi-square distribution with one degree of
freedom and also p-value is obtained from the chi-square distribution. Thus we reject the null
hypothesis, when the probability value is given by
theisandgnificanceoflevelspecifiedathanlessiswhere,11
***
n
i
i
n
i
i xsixp
observed value of the statistic ∆*.
11. APPLICATIONS
In this section, we have used real lifetime data set in length biased Devya distribution and the
model has been compared with Devya distribution.
Data set: Gross and Clark (1975) reported a set of data relating relief in minutes receiving
analgesic of 20 patients. The data is given below:
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.
In order to compare the length baised Devya distribution with Devya distribution, we
consider the criteria like Bayesian information criterion (BIC), Akaike Information Criterion
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1199
(AIC), Akaike Information Criterion Corrected (AICC) and -2 logL. The better distribution is
which corresponds to lesser values of AIC,BIC, AICC and – 2 log L. For calculating AIC,
BIC, AICC and -2 logL can be evaluated by using the formulas as follows:
AIC = 2K - 2logL, BIC = klogn - 2logL, )1(
)1(2
kn
kkAICAICC
Where k is the number of parameters, n is the sample size and -2 logL is the maximized value
of log likelihood function and are showed in table 1.
Table.1: Performance of distributions.
From table 1, we can see that the length biased Devya distribution have the lesser AIC, BIC,
AICC and -2 logL values as compared to Devya distribution. Hence, we can conclude that the
length biased Devya distribution leads to better fit than the Devya distribution.
12. CONCLUSION
A new extension of Devya distribution has been obtained namely the Length Biased
Devya distribution has been introduced to model lifetime data. Its moment generating
function, moments, coefficient of dispersion has been obtained. Other interesting properties
of the distribution such as its hazard rate function, entropies, Bonferroni and Lorenz curves,
have been discussed. The estimation of its parameter have been obtained by using maximum
likelihood estimation. A real lifetime data- set has been presented to show the applications
and performance of Length biased Devya distribution over one parameter Devya
distributions.
Distribution MLE
S.E -2 logL AIC BIC AICC
Devya
Distribution
𝜃
= 1.841946
𝜃 =
=0.1692256
54.50256
56.50436
55.80539
56.72658
Length
biased
Devya
Distribution
𝜃 =2.4679492
𝜃 =0.2123947
46.16638
48.16638
47.48838
48.38860
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1200
REFERENCES
[1]. Fisher, R.A. (1934), The effects of methods of ascertainment upon the estimation of
frequencies, The Annals of Eugenics, 6, 13–25.
[2]. Gross, A.J. and Clark, V.A. (1975): Survival Distribution: Reliability Applications in
the Biometrical Sciences, John Wiley, New York.
[3]. Modi, K. and Gill, V. (2015), Length-biased Weighted Maxwell Distribution,
Pak.j.stat.oper.res. Vol. XI No.4 2015 pp. 465-472.
[4]. Mudasir, S. and Ahmad, S.P. (2018), Characterization and Estimation of the Length
Biased Nakagami Distribution, Pak.j.stat.oper.res. Vol. XIV No.3, pp. 697-715.
[5]. Patil, G. P. and Rao, C. R. (1978). Weighted distributions and size-biased sampling
with applications to wildlife populations and human families, Biometrics, 34, 179-189.
[6]. Rao, C.R. (1965). On discrete distributions arising out of methods of ascertainment,
in Classical and Contagious Discrete Distributions. G.P. Patil, Ed, Statistical Publishing
Society, Calcutta, pp.320-332.
[7]. Rather, A. A. and Subramanian, C. (2019), The Length-Biased Erlang–Truncated
Exponential Distribution with Life Time Data, Journal of Information and
Computational Science, Volume 9, Issue 8, pp. 340-355.
[8]. Shanker, R., (2016): Devya distribution and its applications, International Journal of
Statistics and Applications, 6(4), 189-202.
[9]. Shenbagaraja, R., Rather, A. A. and Subramanian, C. (2019), On Some Aspects of
Length Biased Technique with Real Life Data, Science, Technology and Development,
Volume VIII, Issue IX, pp. 326-335.
Journal of Information and Computational Science
Volume 9 Issue 11 - 2019
ISSN: 1548-7741
www.joics.org1201