Research Article Volume 6 Issue 1
Department of Statistics, Eritrea Institute of Technology, Eritrea
Correspondence: Rama Shanker, Department of Statistics, Eritrea Institute of Technology, Asmara, Eritrea
Received: June 03, 2017 | Published: June 13, 2017
Citation: Shanker R. A quasi shanker distribution and its applications. Biom Biostat Int J. 2017;6(1):267-276. DOI: 10.15406/bbij.2017.06.00156
In the present paper, a two-parameter quasi Shanker distribution (QSD) which includes one parameter Shanker distribution introduced by Shanker1 as a special case has been proposed. Its statistical and mathematical properties including moments and moments based measures, hazard rate function, mean residual life function, stochastic ordering, mean deviations, Bonferroni and Lorenz curves and stress-strengthreliability have also been discussed. The method of maximum likelihood estimation has been discussed for estimating the parameters of QSD. Finally, the goodness of fit of the QSD has been discussed with two real lifetime data and the fit is quite satisfactory over one parameter exponential, Lindley and Shanker distributions.
Keywords: shanker distribution, moments, hazard rate function, mean residual life function, stochastic ordering, mean deviations, stress-strength reliability, estimation of parameters, goodness of fit
Shanker1 has introduced a one parameter lifetime distribution for modeling lifetime data from biomedical science and engineering having probability density function(pdf) and cumulative distribution function(cdf) given by
f1(x;θ)=θ2θ2+1(θ+x)e−θx ;x>0, θ>0
…. (1.1)
F1(x,θ)=1−[1+θxθ2+1]e−θx ;x>0,θ>0
(1.2)
Shanker1 has shown that it gives better fit than both one parameter exponential and Lindley2 distributions. This distribution is a mixture of exponential (θ)
and gamma (2,θ)
distributions with their mixing proportionθ2θ2+1
and1θ2+1
respectively.
The first four moments about origin of Shanker distribution obtained by Shanker1 are given as
μ1′=θ2+2θ(θ2+1) ,μ2′=2(θ2+3)θ2(θ2+1) , μ3′=6(θ2+4)θ3(θ2+1) ,μ4′=24(θ2+5)θ4(θ2+1)
The central moments of Shanker distribution obtained by Shanker1 are
μ2=θ4+4θ2+2θ2(θ2+1)2
μ3=2(θ6+6θ4+6θ2+2)θ3(θ2+1)3
μ4=3(3θ8+24θ6+44θ4+32θ2+8)θ4(θ2+1)4
Shanker1 studied its important properties including coefficient of variation, skewness, kurtosis, Index of dispersion, hazard rate function, mean residual life function, stochastic ordering, mean deviations, order statistics, Bonferroni and Lorenz curves, Renyi entropy measure, and stress-strength reliability. The discrete Poisson - Shanker distribution, a Poisson mixture of Shanker distribution has also been studied by Shanker.3.
Recall that the Lindley distribution, introduced by Lindley2 in the context of Bayesian analysis as a counter example of fiducial statistics, is defined by its pdf and cdf
f2(x;θ)=θ2θ+1(1+x)e−θx ;x>0, θ>0
(1.3)
F2(x;θ)=1−[1+θxθ+1]e−θx ;x>0, θ>0
(1.4)
In this paper, a two - parameter quasi Shanker distribution (QSD), of which one parameter Shanker distribution introduced by Shanker1 is a particular case, has been proposed. Its raw moments and central moments have been obtained and coefficients of variation, skewness, kurtosis and index of dispersion have been discussed. Some of its important mathematical and statistical properties including hazard rate function, mean residual life function, stochastic ordering, mean deviations, Bonferroni and Lorenz curves and stress-strength reliability have also been discussed. The estimation of the parameters has been discussed using maximum likelihood estimation. The goodness of fit of QSD has been illustrated with two real lifetime data sets and the fit has been compared with one parameter exponential, Lindley and Shanker distributions.
A two - parameter quasi Shanker distribution (QSD) having parameters θ and α is defined by its pdf
f(x;θ,α)=θ3θ3+θ+2α(θ+x+α x2) e−θ x ;x>0, θ>0 , θ3+θ+2α>0.
(2.1)
It can be easily verified that (2.1) reduces to the Shanker distribution (1.1) at α=0
. It can be easily verified that QSD is a three-component mixture of exponential(θ)
, gamma (2,θ)
and gamma(3,θ)
distributions. We have
f(x;θ,α)=p1 f1(x;θ)+p2f2(x;2,θ)+(1−p1−p2)f3(x;3,θ) (2.2)
where
p1=θ3θ3+θ+2α , p2=θθ3+θ+2α
,
f1(x;θ)=θ e−θ x; x>0, θ>0
f2(x;2,θ)=θ2Γ(2) e−θ xx2−1; x>0, θ>0
f3(x;3,θ)=θ3Γ(3) e−θ xx3−1; x>0, θ>0
The corresponding cdf of QSD (2.1) can be obtained as
F(x;θ,α)=1−[1+α θ2x2+θ x(θ+2α)θ3+θ+2α]e−θ x; x>0,θ>0 (2.3)
The nature and behavior of the pdf and the cdf of QSD for varying values of the parameters θ and α have been explained graphically and presented in Figures 1 & 2, respectively.
The r th moment about origin of QSD can be obtained as
μr′=r![θ3+(r+1)θ+(r+1)(r+2)α]θr(θ3+θ+2α) ;r=1,2,3,.. (3.1)
Thus, the first four moments about origin of QSD are given by
μ1′=θ3+2θ+6αθ(θ3+θ+2α)
, μ2′=2(θ3+3θ+12α)θ2(θ3+θ+2α)
μ3′=6(θ3+4θ+20α)θ3(θ3+θ+2α)
, μ4′=24(θ3+5θ+30α)θ4(θ3+θ+2α)
Using relationship between central moments and moments about origin, the central moments of QSD (2.1) are thus obtained as
μ2=θ6+4θ4+16 θ3α+2θ2+12 θ α+12α2θ2(θ3+θ+2α)2
μ3=2{θ9+6θ7+30θ6α+6 θ5+42θ4α+(36α2+2)θ3+18θ2α+36θ α2+24α3}θ3(θ3+θ+2α)3
μ4=3{3θ12+24θ10+128θ9α+44θ8+344θ7α+(408α2+32)θ6+320 θ5α+(768α2+8)θ4+(576α3+96α)θ3+336θ2α2+480 θα3+240α4}θ4(θ3+θ+2α)4
The coefficient of variation(C.V)
, coefficient of skewness(√β1)
, coefficient of kurtosis (β2)
and index of dispersion (γ)
of QSD are obtained as
C.V=σμ1′=√θ6+4θ4+16 θ3α+2θ2+12 θ α+12α2θ3+2θ+6α
√β1=μ3μ23/2=2{θ9+6θ7+30θ6α+6 θ5+42θ4α+(36α2+2)θ3+18θ2α+36θ α2+24α3}(θ6+4θ4+16 θ3α+2θ2+12 θ α+12α2)3/2
β2=μ4μ22=3{3θ12+24θ10+128θ9α+44θ8+344θ7α+(408α2+32)θ6+320 θ5α+(768α2+8)θ4+(576α3+96α)θ3+336θ2α2+480 θα3+240α4}(θ6+4θ4+16 θ3α+2θ2+12 θ α+12α2)2
γ=σ2μ1′=θ6+4θ4+16 θ3α+2θ2+12 θ α+12α2θ(θ3+θ+2α)(θ3+2θ+6α)
Graphs of C.V,√β 1 ,β 2 and γ of QSD for varying values of the parameters θ andα have been presented in Figure 3.
SupposeX be a continuous random variable with pdf f(x) and cdf 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
h(x)=limΔx→0P(X<x+Δx | X>x)Δx=f(x)1−F(x)
(4.1)
And m(x)=E[X−x|X>x] = 11−F(x)∫∞x[1−F(t)] dt
(4.2)
The corresponding hazard rate functionh(x)
, and the mean residual life functionm(x)
of QSD are thus obtained as
h(x)=θ3(θ+x+αx2)α θ2x2+θ(θ+2α)x+(θ3+θ+2α)
(4.3)
and m(x)=1[α θ2x2+θ(θ+2α)x+(θ3+θ+2α)]e−θx∞∫x[α θ2t2+θ(θ+2α)t+(θ3+θ+2α)]e−θ tdt
=α θ2x2+θ(θ+4α)x+(θ3+2θ+6α)θ[α θ2x2+θ(θ+2α)x+(θ3+θ+2α)]
(4.4)
It can be easily verified that h(0)=θ4θ3+θ+2α=f(0)
andm(0)=θ3+2θ+6αθ(θ3+θ+2α)=μ1′
The nature and behavior of h(x)
and m(x)
of QSD for varying values of parameters θ
and α
have been shown graphically in Figures 4 & 5. It is obvious that h(x)
of QSD is monotonically increasing whereas h(x)
is monotonically decreasing
Stochastic ordering of positive continuous random variables is an important tool for judging their comparative behavior. A random variableX is said to be smaller than a random variable Y in the
The following results due to Shaked and Shanthikumar4 are well known for establishing stochastic ordering of distributions
X≤lrY⇒X≤hrY⇒X≤mrlY
⇓X≤stY
The QSD is ordered with respect to the strongest ‘likelihood ratio ordering’ as shown in the following theorem:
Theorem: LetX
∼
QSD(θ1,α1)
and Y
∼
QSD(θ2,α2)
. Ifα1=α2 and θ1>θ2
(orθ1=θ2 and α1<α2
), thenX≤lrY
and henceX≤hrY
,X≤mrlY
andX≤stY
.
Proof: We have
fX(x;θ1,α1)fY(x;θ2,α2)=θ13(θ23+ θ2+2α2)θ23(θ13+ θ1+2α1) (θ1+x+α1x2θ2+x+α2x2) e−(θ1−θ2) x ;x>0
Now
lnfX(x;θ1,α1)fY(x;θ2,α2)=log[θ13(θ23+ θ2+2α2)θ23(θ13+ θ1+2α1)]+ln(θ1+x+α1x2θ2+x+α2x2)−(θ1−θ2)x
This gives
ddx{lnfX(x;θ1,α1)fY(x;θ2,α2)}=(θ2−θ1)+(α2−α1)+2(α1θ2−α2θ1)x+2(α1−α2)x2(θ1+x+α1x2)(θ2+x+α2x2)−(θ1−θ2)
Thus ifα1=α2 and θ1>θ2 orθ1=θ2 and α1<α2 ,ddxlnfX(x;θ1,α1)fY(x;θ2,α2)<0 . This means thatX≤lrY and henceX≤hrY ,X≤mrlY andX≤stY .
The amount of scatter in a population is measured to some extent by the totality of deviations usually from mean and median. These are known as the mean deviation about the mean and the mean deviation about the median defined by
δ1(X)=∞∫0|x−μ| f(x)dx
andδ2(X)=∞∫0|x−M| f(x)dx
, respectively, where μ=E(X)
andM=Median (X)
. The measures δ1(X)
andδ2(X)
can be calculated using the following simplified relationships
δ1(X)=μ∫0(μ−x)f(x)dx+∞∫μ(x−μ)f(x)dx
=μF(μ)−μ∫0x f(x)dx−μ[1−F(μ)]+∞∫μx f(x)dx
=2μF(μ)−2μ+2∞∫μx f(x)dx
=2μF(μ)−2μ∫0x f(x)dx
(6.1)
and
δ2(X)=M∫0(M−x)f(x)dx+∞∫M(x−M)f(x)dx
=M F(M)−M∫0x f(x)dx−M[1−F(M)]+∞∫Mx f(x)dx
=−μ+2∞∫Mx f(x)dx
=μ−2M∫0x f(x)dx
(6.2)
Using p.d.f. (2.1) and expression for the mean of QSD, we get
μ∫0x f(x)dx=μ−{α θ3μ3+θ2(θ+3α)μ2+θ(θ3+2θ+6α)μ+(θ3+2θ+6α)}e−θ μθ(θ3+θ+2α)
(6.3)
M∫0x f(x)dx=μ−{α θ3M3+θ2(θ+3α)M2+θ(θ3+2θ+6α)M+(θ3+2θ+6α)}e−θ Mθ(θ3+θ+2α)
(6.4)
Using expressions from (6.1), (6.2), (6.3), and (6.4), the mean deviation about mean, δ1(X)
and the mean deviation about median, δ2(X)
of QSD are finally obtained as
δ1(X)=2{α θ2μ2+θ(θ+4 α) μ+(θ3+2θ+6α)}e−θ μθ(θ3+θ+2α)
(6.5)
δ2(X)=2{α θ3M3+θ2(θ+3α)M2+θ(θ3+2θ+6α)M+(θ3+2θ+6α)}e−θ Mθ(θ3+θ+2α)−μ
(6.6)
The Bonferroni and Lorenz curves5 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
B(p)=1pμq∫0x f(x) dx=1pμ[∞∫0x f(x)dx−∞∫qx f(x) dx]=1pμ[μ−∞∫qx f(x) dx]
(7.1)
and L(p)=1μq∫0x f(x) dx=1μ[∞∫0x f(x)dx−∞∫qx f(x) dx]=1μ[μ−∞∫qx f(x) dx]
(7.2)
Respectively or equivalently
B(p)=1pμp∫0F−1(x) dx
(7.3)
and L(p)=1μp∫0F−1(x) dx
(7.4)
Respectively, where μ=E(X)
and q=F−1(p)
.
The Bonferroni and Gini indices are thus defined as
B=1−1∫0B(p) dp
(7.5)
and G=1−21∫0L(p) dp
(7.6) respectively.
Using p.d.f. of QSD (2.1), we get
∞∫qx f(x) dx={α θ3q3+θ2(θ+3α)q2+θ(θ3+2θ+6α)q+(θ3+2θ+6α)}e−θ qθ(θ3+θ+2α)
(7.7)
Now using equation (7.7) in (7.1) and (7.2), we get
B(p)=1p[1−{α θ3q3+θ2(θ+3α)q2+θ(θ3+2θ+6α)q+(θ3+2θ+6α)}e−θ qθ3+2θ+6α]
(7.8)
and
L(p)=1−{α θ3q3+θ2(θ+3α)q2+θ(θ3+2θ+6α)q+(θ3+2θ+6α)}e−θ qθ3+2θ+6α (7.9)
Now using equations (7.8) and (7.9) in (7.5) and (7.6), the Bonferroni and Gini indices of QSD are thus obtained as
B=1−{α θ3q3+θ2(θ+3α)q2+θ(θ3+2θ+6α)q+(θ3+2θ+6α)}e−θ qθ3+2θ+6α
(7.10)
G=2{α θ3q3+θ2(θ+3α)q2+θ(θ3+2θ+6α)q+(θ3+2θ+6α)}e−θ qθ3+2θ+6α−1
(7.11)
The stress- strength reliability describes the life of a component which has random strengthX that is subjected to a random stressY . When the stress applied to it exceeds the strength, the component fails instantly and the component will function satisfactorily till X>Y . Therefore,R=P(Y<X) is a measure of component reliability and in statistical literature it is known as stress-strength parameter. It has wide applications in almost all areas of knowledge especially in engineering such as structures, deterioration of rocket motors, static fatigue of ceramic components, aging of concrete pressure vessels etc. Let X and Y be independent strength and stress random variables having QSD (2.1) with parameter (θ1,α1) and(θ2,α2) respectively. Then the stress-strength reliability R of QSD (2.1) can be obtained as
R=P(Y<X)=∞∫0P(Y<X|X=x)fX(x)dx
=∞∫0f(x;θ1,α1) F(x;θ2,α2)dx
=1−θ13[θ1 θ27+(4θ12+1)θ26+(6 θ13+5 θ1+2α1)θ25+(4θ14+10θ12+4α1θ1+4α2θ1+3)θ24+(θ15+10θ13+2α1θ12+14 α2θ12+8α1+7θ1+2α2θ1+6α2)θ23+(5θ14+18 α2θ13+4α2θ12+5θ12+16α1α2+10α1θ1+14α2θ2+6α2)θ22+(θ15+10α2θ14+2α2θ13+θ13+10α2θ12+2α1θ12+20α1α2θ1+24α1α2+6α2θ1)θ2+2(α2θ15+2α1α2θ12+2α2θ13)](θ13+θ1+2α1)(θ23+θ2+2α2)(θ1+θ2)5
.
It can be easily verified that at α1=0 and α2=0 , the above expression reduces to the corresponding expression for Shanker distribution introduced by Shanker.1
Let (x1, x2, x3, ... ,xn) be a random sample of size n from QSD (2.1)). The likelihood function, L of (2.1) is given by
L=(θ3θ3+θ+2α)nn∏i=1(θ+xi+α xi2) e−n θ ˉx
The natural log likelihood function is thus obtained as
lnL=nln(θ3θ3+θ+2α)+n∑i=1ln(θ+xi+α xi2)−n θ ˉx
The maximum likelihood estimates (MLE) (ˆθ,ˆα) of (θ,α) are then the solutions of the following non-linear equations
∂lnL∂θ=3nθ−n (3θ2+1)θ3+θ+2α+n∑i=11θ+xi+α xi2−n ˉx=0
∂lnL∂α=−2nθ3+θ+2α+n∑i=1xi2θ+xi+α xi2=0
where ˉx is the sample mean.
These two natural log likelihood equations do not seem to be solved directly because they are not in closed forms. However, the Fisher’s scoring method can be applied to solve these equations. For, we have
∂2lnL∂θ2=−3nθ2+n(3θ4−6θ3+5θ2−12θ α+1)α2(θ3+θ+2α)2−n∑i=11(θ+xi+α xi2)2
∂2lnL∂θ ∂α=2n (3θ2+1)(θ3+θ+2α)2−n∑i=1xi2(θ+xi+α xi2)2
∂2lnL∂α2=4n(θ3+θ+2α)2−n∑i=1xi4(θ+xi+α xi2)2
The solution of following equations gives MLE’s (ˆθ,ˆα) of (θ,α) of QSD
[∂2lnL∂θ2∂2lnL∂θ ∂α∂2lnL∂θ ∂α∂2lnL∂α2]ˆθ=θ0ˆα=α0[ˆθ−θ0ˆα−α0]=[∂lnL∂θ∂lnL∂α]ˆθ=θ0ˆα=α0
where θ0 and α0 are the initial values of θ and α , respectively. These equations are solved iteratively till sufficiently close values of ˆθ and ˆα are obtained.
In this section, the goodness of fit of QSD has been discussed with two real lifetime data sets from engineering and the fit has been compared with one parameter exponential, Lindley and Shanker distributions. The following two data sets have been considered.
Data set 1
This data set is the strength data of glass of the aircraft window reported by Fuller et al.6
18.83 | 20.8 | 21.657 | 23.03 | 23.23 | 24.05 | 24.321 | 25.5 | 25.52 | 25.8 | 26.69 | 26.77 | 26.78 |
27.05 | 27.67 | 29.9 | 31.11 | 33.2 | 33.73 | 33.76 | 33.89 | 34.76 | 35.75 | 35.91 | 36.98 | 37.08 |
37.09 | 39.58 | 44.045 | 45.29 | 45.381 |
Data set 2
The following data represent the tensile strength, measured in GPa, of 69 carbon fibers tested under tension at gauge lengths of 20mm, Bader and Priest.7
1.312 | 1.314 | 1.479 | 1.552 | 1.7 | 1.803 | 1.861 | 1.865 | 1.944 | 1.958 | 1.966 | 1.997 | 2.006 |
2.021 | 2.027 | 2.055 | 2.063 | 2.098 | 2.14 | 2.179 | 2.224 | 2.24 | 2.253 | 2.27 | 2.272 | |
2.274 | 2.301 | 2.301 | 2.359 | 2.382 | 2.382 | 2.426 | 2.434 | 2.435 | 2.478 | 2.49 | 2.511 | |
2.514 | 2.535 | 2.554 | 2.566 | 2.57 | 2.586 | 2.629 | 2.633 | 2.642 | 2.648 | 2.684 | 2.697 | |
2.726 | 2.77 | 2.773 | 2.8 | 2.809 | 2.818 | 2.821 | 2.848 | 2.88 | 2.954 | 3.012 | 3.067 | |
3.084 | 3.09 | 3.096 | 3.128 | 3.233 | 3.433 | 3.585 | 3.585 |
In order to compare the considered distributions, values of −2lnL , AIC(Akaike Information Criterion) and K-S Statistic ( Kolmogorov-Smirnov Statistic) for the data sets have been computed and presented in Table 1. The formula for AIC and K-S Statistic is defined as follow:
AIC=−2lnL+2k and K-S=Supx|Fn(x)−F0(x)| , where k= number of parameters, n= sample size, Fn(x) is the empirical distribution function and F0(x) is the theoretical cumulative distribution function.. The best distribution corresponds to lower values of−2lnL , AIC and K-S statistic. It can be easily seen from table 1 that the QSD gives better fit than one parameter exponential, Lindley and Shanker distributions and hence it can be considered as an important distribution for modeling lifetime data from engineering.
Data sets |
Distributions |
ML estimates |
Standard errors |
−2lnL |
AIC |
K-S statistic |
1 |
QSD |
ˆθ=0.097330 |
0.0101017 |
240.53 |
244.53 |
0.298 |
ˆα=13.623065 |
52.81378 |
|||||
Shanker |
ˆθ=0.6471636 |
0.0082 |
252.35 |
254.35 |
0.358 |
|
Lindley |
ˆθ=0.062990 |
0.008 |
253.98 |
255.98 |
0.365 |
|
Exponential |
ˆθ=0.032449 |
0.005822 |
274.53 |
276.53 |
0.458 |
|
2 |
QSD |
ˆθ=1.20552 |
0.083861 |
186.78 |
190.78 |
0.314 |
ˆα=49.73844 |
34.58363 |
|||||
Shanker |
ˆθ=0.658030 |
0.052373 |
233 |
235 |
0.369 |
|
Lindley |
ˆθ=0.65450 |
0.058031 |
238.38 |
240.38 |
0.401 |
|
Exponential |
ˆθ=0.407942 |
0.04911 |
261.73 |
263.73 |
0.448 |
Table 1 MLE’s, −2 ln L , standard error, AIC, and K-S statistic of the fitted distributions of data sets 1 and 2
A two-parameter quasi Shanker distribution (QSD), of which one parameter Shanker distribution introduced by Shanker1 is a particular case, has been suggested and investigated. Its mathematical properties including moments, coefficient of variation, skewness, kurtosis, index of dispersion, hazard rate function, mean residual life function, stochastic ordering, mean deviations, Bonferroni and Lorenz curves, and stress-strength reliability have been discussed. For estimating its parameters method of maximum likelihood estimation has been discussed. Finally, two numerical examples of real lifetime data sets has been presented to test the goodness of fit of QSD over exponential, Lindley and Shanker distributions and the fit by QSD has been quite satisfactory. Therefore, QSD can be recommended as an important two-parameter lifetime distribution.
None.
Authors declare that there are no conflicts of interests.
©2017 Shanker, et al. This is an open access article distributed under the terms of the, which permits unrestricted use, distribution, and build upon your work non-commercially.
2 7