Journal of Statistical Theory and Applications

Volume 17, Issue 3, September 2018, Pages 554 - 571

The Exponential Lindley Odd Log-Logistic-G Family: Properties, Characterizations and Applications

Authors
Mustafa Ç. Korkmaz*, mcagatay@artvin.edu.tr mustafacagataykorkmaz@gmail.com
Department of Measurement and Evaluation, Artvin Çoruh University, Artvin, TURKEY
Haitham M. Yousofhaitham.yousof@fcom.bu.edu.eg
Department of Statistics, Mathematics and Insurance, Benha University, Benha, EGYPT
G. G. Hamedanigholamhoss.hamedani@marquette.edu
Department of Mathematics, Statistics and Computer Science, Marquette University, USA
*Corresponding author
Corresponding Author
Received 3 May 2017, Accepted 28 August 2017, Available Online 30 September 2018.
DOI
10.2991/jsta.2018.17.3.10How to use a DOI?
Keywords
Exponential Lindley family; odd log-logistic family; mixture distribution family; maximum likelihood; moments; characterizations; generalized distribution family; mixture family
Abstract

A new family of distributions called the exponential Lindley odd log-logistic G family is introduced and studied. The new generator generalizes three newly defined G families and also defines two new G families. We provide some mathematical properties of the new family. Characterizations based on truncated moments as well as in terms of the hazard function are presented. The maximum likelihood is used for estimating the model parameters. We assess the performance of the maximum likelihood estimators in terms of biases and mean squared errors by means of a simulation study. Finally, the usefulness of the family is illustrated by means of three real data sets. The new model provides consistently better fits than other competitive models for these data sets.

Copyright
© 2018, the Authors. Published by Atlantis Press.
Open Access
This is an open access article under the CC BY-NC license (http://creativecommons.org/licences/by-nc/4.0/).

1. Introduction

Recently, several families of continuous uni-variate distributions have been constructed by extending common families of continuous models. These generalized distributions give more flexibility by adding one or more parameters to the baseline model. Many classes can be cited such as the Marshall-Olkin-G family by Marshall and Olkin [25], transmuted exponentiated generalized-G family by Yousof et al. [34], Burr X-G by Yousof et al. [35], type I half-logistic family by Cordeiro et al. [12], Zografos-Balakrishnan odd log-logistic family of distributions by Cordeiro et al. [13], a new generalized two-sided family of distributions by Korkmaz and Genç [22], generalized odd log-logistic family by Cordeiro et al. [10], odd-Burr generalized family by Alizadeh et al. [4], beta Weibull G by Yousof et al. [36], exponentiated generalized-G Poisson family by Aryal and Yousof [8], type I general exponential class by Hamedani et al. [20] and beta transmuted-H by Afify et al. [2] among others.

Recently, Gleaton and Lynch [17] introduced a class of distributions called the odd log-logistic family with one extra shape parameter α > 0 with cumulative distribution function (cdf)

H(x;ξ)=G(x;ξ)αG(x;ξ)α+G¯(x;ξ)α,
where G(x; ξ) is the baseline cdf depending on the vector parameter ξ and G¯(x;ξ)=1-G(x;ξ) and ξ the parameter vector of associated parameters from baseline distribution. Alzaatreh et al. [5] defined the TX family with cdf
F(x)=aW[G(x;ξ)]r(t)dt,
where r(t) is the probability density function (pdf) of the random variable of T ∈ [a, b] for −∞ < a < b < ∞, G(x; ξ) is cdf of the baseline random variable X and W [G(x)] is a function of the baseline cdf which satisfies the following conditions: i)W [G(x)] ∈ [a, b], ii)W [G(x)] is differentiable and monotonically non-decreasing, iii)limx→−∞W [G(x)] = a and limx→∞W [G(x)] = b. The pdf of the TX family is given by
f(x)={xW[G(x;ξ)]}r(W[G(x;ξ)]).
On the other hand, an extension of the Lindley distribution, denoted by exponential Lindley (EL), was introduced by Gomez et al. [18]. The cdf and pdf of this extension are given, respectively, by
GEL(x;β,θ)=1-(θ+β+θβx)θ+βe-θx
and
gEL(x;β,θ)=θ2(1+βx)θ+βe-θx
where x > 0, β > 0 and θ > 0. Clearly, distribution of this extension is exponential distribution for β = 0 and is Lindley distribution for β = 1. Also, this distribution is more flexible than exponential and Lindley distributions.

The goal of this paper is to introduce a new flexible and wider family of the distributions based on T–X family using the EL as the generator. The paper is organized as follows. The new family is introduced in Section 2. Some members and sub-families of the new family are introduced in Section 3. In Section 4, the series expansions for cdf and pdf of the new family are presented. In Section 5, some of its mathematical properties are derived. Section 6 deals with some characterizations of the new family. In Section 7, the maximum likelihood method is presented. In Section 8, a simulation study is performed to evaluate the efficiency of the Maximum Likelihood method. In Section 9, we illustrate the importance of the new family by means of three applications to real data sets. The paper is concluded in Section 10.

2. The new family

We define the new family by taking W [G(x; ξ)] = −log[1 − H(x; ξ)] and r(t) = gEL (t; β, θ). From (1.1), (1.2), (1.3) and (1.4), we introduce cdf and pdf of the new family as

F(x,α,β,θ,ξ)=1-G¯(x)αθ(θ+β)[G(x)α+G¯(x)α]θ{θ+β-θβlog[G¯(x)αG(x)α+G¯(x)α]}
and
f(x,α,β,θ,ξ)=αθ2θ+βg(x)G¯(x)αθ-1G(x)α-1[G(x)α+G¯(x)α]θ-1{1-βlog[G¯(x)αG(x)α+G¯(x)α]}
respectively, where G(x) = G(x; ξ) is the baseline cdf, g(x) = g(x; ξ)is the baseline pdf, α > 0, β > 0 and θ > 0 are three extra parameters. We call it exponential Lindley odd log-logistic family and denote it by ELOLL−G(α, β, θ, ξ). Further, the ELOLL-G family is in fact a mixture of the exponential-odd log-logistic-G family (EOLL), whose cdf is 1-(G¯(x)αG¯(x)α+G(x)α)θ and gamma(2, θ)-odd log-logistic-G family (GaOLL), whose cdf is 1-[1-log(G¯(x)αG¯(x)α+G(x)α)θ](G¯(x)αG¯(x)α+G(x)α)θ, with mixing rate θ/(θ + β)and β/(θ + β) respectively. So, we express the ELOLL-G family as
ELOLL-G(α,β,θ,ξ)=θθ+βEOLL-G(α,θ,ξ)+βθ+βGaOLL-G(α,θ,ξ).
As a result, we can say that the ELOLL-G family is a mixture family. The hazard rate function (hrf) of this family is given by,
h(x,α,β,θ,ξ)=αθ2g(x)G(x)α-1(1-βlog[G¯(x)αG(x)α+G¯(x)α])(G¯(x)[G(x)α+G¯(x)α])(θ+β-θβlog[G¯(x)αG(x)α+G¯(x)α]).
The quantile function (qf) of ELOLL-G is given as follows: if U has a uniform U(0, 1) the solution of the non-linear equation x=Q(U)=QG(TU1/α[TU1/α+(1-TU)1/α]-1) has random number on the ELOLL−G(α, β, θ, ξ), where QG(·) = G−1(·) is the qf of the baseline distribution, QEL(u) is the solution of (1.3) that is the qf of the EL distribution and TU = (1 − e−QEL(U)). Hence, if U is a uniform random variable on U(0,1), then X = Q(U) has the ELOLL-G distribution. Also, the mixture representation in (2.3) can be used to generate random number from the ELOLL-G distribution.

3. Some special cases of the ELOLL-G family

The ELOLL-G family includes important sub-families. We mention these sub-families in Table (1). Also, the (2.1) and (2.2) will be most tractable when g(x) and G(x) have simple analytic forms. Here, we provide three special models of the ELOLL-G family. These special models generalize some well-known distributions reported in the literature. We note that since the cdf, pdf and hrf of any ELOLL-G distribution will be easily determined by (2.1), (2.2) and (2.4), we dont give them in following sub-sections.

α β θ Sub-family References
1 Exponential Lindley-G New
1 1 Lindley-G Çakmakyapan and Ozel [9]
1 0 Lehmann Type II-G or Proportional hazard rate model Gupta et al. [19]
0 Odd Burr-G or Lehmann Type II odd log-logistic-G Alizadeh et al. [4]
1 Lindley odd log-logistic-G New
1 0 1 G(x) -
Table 1:

Sub families of the ELOLL-G family.

3.1. The ELOLL-normal distribution

We define the ELOLL-normal (ELOLL-N) distribution from (2.2) by taking G(x;μ,σ)=Φ(x-μσ) and g(x;μ,σ)=σ-1ϕ(x-μσ) with ξ = (µ ∈ ℜ, σ > 0), where ϕ(·) and Φ(·) are the pdf and cdf of the standard normal distribution, respectively. Its pdf is denoted by ELOLL−N(Θ) where Θ = (α, β, θ, µ, σ). We plot this pdf and its hrf in Figure 1. From Figure 1, we see that the pdf shapes of the ELOLL-N are skewed and bi-modal. Also, its hrf are increasing or firstly unimodal and then increasing. The ordinary normal distribution acquires these properties.

Fig. 1:

Plots of the pdf and hrf of the ELOLL-N distributions

3.2. The ELOLL-Weibull distribution

We now consider the Weibull distribution as a baseline distribution with pdf g(x; λ, γ) = γλγxγ−1e−(λx)γ and cdf G(x; λ, γ) = 1−e(λx)γ for x, λ, γ > 0. Its pdf is denoted by ELOLL−W(Θ) where Θ = (α, β, θ, λ, γ). Its pdf and hrf for selected parameter values are displayed in Figure 2. From Figure 2, we see that pdf and hrf of the ELOLL-W model have various shapes. The pdf shapes are decreasing, unimodal, bi-modal, firstly decreasing then unimodal shaped, U-shaped. Also, its hrf shapes are increasing, decreasing, unimodal, bathtube-shaped and firstly unimodal and then increasing shaped. So, we can say that ELOLL-W distribution can be useful for modelling data.

Fig. 2:

Plots of the pdf and hrf of the ELOLL-W distributions

3.3. The ELOLL-Lomax distribution

The pdf and cdf of the Lomax distribution with scale parameter λ > 0 and shape parameter k > 0 are given by g(x,λ,k)=kλ(1+xλ)-k-1 and G(x,λ,k)=1-(1+xλ)-k respectively. Its pdf is denoted by ELOLL−Lx(Θ) where Θ = (α, β, θ, λ, k). Plots of its pdf and hrf for selected parameter values are displayed in Figure 3.

Fig. 3:

Plots of the pdf and hrf of the ELOLL-Lx distributions

4. Expansions for cdf and pdf

In this Section, we will obtain cdf and pdf expansions of the ELOLL-G family using Taylor series of log(1-z)=-k=0zk+1(k+1),|z|<1 and others series expansions. Then equation (2.1) can be written as

F(x)=1-G¯(x)αθA[G(x)α+G¯(x)α]θB-i=0θβj=0(-1)j(αθj)G(x)α(i+1)+jD(θ+β)(i+1)[G(x)α+G¯(x)α]θ+i+1C,
where the quantities A, B, C and D are defined in Appendix A . The last equation can be written as
F(x)=1-[k=0dkG(x)kk=0bkG(x)k+i=0θβk=0pkG(x)k(θ+β)(i+1)k=0hkG(x)k],
where bk and hk are defined in the Appendix A. Then
F(x)=1-[k=0ϒkG(x)k+i=0θβ(θ+β)(i+1)k=0ΩkG(x)k],
where ϒ0=d0b0 and Ω0=p0h0 for k ≥ 1 and
ϒk=1b0[dk+1b0r=1kbrϒk-r]andΩk=1h0[pk+1h0r=1khrΩk-r],
then
F(x)=1-k=0ζkΠk(x)=k=0vkΠk(x),
where Ψk=i=0θβ(θ+β)(i+1)Ωk, υ0 = 1 − ζk = 1 kk) and for k ≥ 1 we have υ0 = −ζk and Πk(x) is the cdf of the exponentiated-G (exp-G) family with power parameter (γ). Equation (4.1) reveals that the density of ELOLL-G family is a linear combination of exp-G densities. So, some mathematical properties of this family can be determined from those of the exp-G distribution. By differentiating (4.1), we obtain the same mixture representation
f(x)=k=0υk+1πk+1(x),
where πγ(x) = γg(x)G(x)γ−1 represents the exponentiated-G family density with power parameter γ > 0. The equations (4.1) and (4.2) are the main results of this section.

5. Properties

The rth moment of X, say µ′r, follows from (4.2) as

μr=E(Xr)=k=0υk+1E(Yk+1r).
Henceforth, Yγ denotes the exp-G distribution with power parameter (γ). For γ > 0, we have E(Yγr)=γ-xrg(x;ξ)G(x;ξ)γ-1dx, which can be computed numerically in terms of the baseline quantile function (qf) QG (u; ξ) = G−1 (u; ξ) as E(Yγr)=γ01QG(u;ξ)ruγ-1du. The nth central moment of X, say Mn, is given by
Mn=E(X-μ1)n=r=0n(nr)(-μ1)n-rE(Xr),=r=0nk=0υk+1(nr)(nr)(-μ1)n-rE(Xr).
Here, we provide two formulae for the mgf MX (t) = E(etX) of X. Clearly, the first one can be derived from equation (5) as MX(t)=k=0vk+1Mk+1(t), where Mk+1 (t) is the mgf of Yk+1. Hence, MX (t) can be determined from the exp-G generating function. A second formula for MX (t) follows from (4.2) as MX(t)=k=0υk+1ς(t,k), where ς(t,k)=01exp[tQG(u)]ukdu and QG(u) is the qf corresponding to G(x; φ), i.e., QG(u) = G−1(u; φ). For the ELOLL-W model we get
μr=E(Xr)=k,h=0υk+1k+1λrbh(k+1)Γ(1+rγ),r>-γwithbh(d)=(-1)h(h+1)r+γγ(d-1h),d>0.

Using the series expansion (1-z)a=i=0(ai)(-z)i for |z| < 1, one can expand (k+1)βαβxβ-1exp[-(αx)β]{1-exp[-(αx)β]}k as

f(x)=(k+1)h=0(kh)(-1)hh+1f(h+1)1/γλ(x),
where f(h+1)1/γλ(·) denotes the pdf of a two-parameter Weibull distribution with λ replaced by (h + 1)1 λ. So, whenever possible, (5.1) can be used to derive moment generating function of the ELOLL-W model from those of a two-parameter Weibull distribution. Consider pΨq(·) as the complex parameter Wright generalized hypergeometric function with p numerator and q denominator parameters (Kilbas et al. [21], Equation (1.9)) defined by the series
Ψp[(α1,A1),,(αp,Ap)(β1,B1),,(βq,Bq);z]=n=0j=1pΓ(αj+Ajn)j=1qΓ(βj+Bjn)znn!,forzβ
where αj, βkBk, Aj, Bk ≠ 0, j=1,p¯, k=1,q¯ and the series converges for 1+j=1qBj-j=1pAj>0, compare with Mathai and Saxena [26] and Srivastava et al. [31]. This function was originally introduced by Wright [33]. Let X be a random variable having the pdf (3.2). Following similar algebraic developments of Nadarajah et al. [28], we can write the moment generating function (mgf) of the ELOLL-W model
MX(t)=k,h=0υk+1(-1)hh+11Ψ0[(1,1/γ)-;t{(h+1)1/γλ}-1].
Hypergeometric functions are included as in-built functions in most popular algebraic mathematical software packages, so the special function in (5.2) and hence (5.3) can be easily evaluated by the software packages Maple, Matlab and Mathematica using known procedures.

Suppose X1,...,Xn is a random sample from an ELOLL-G distribution. Let Xi:n denote the ith order statistic. The pdf of Xi:n can be expressed as

fi:n(x)=f(x)B(i,n-i+1)j=0n-i(-1)j(n-ij)F(x)j+i-1.
Following similar algebraic developments of Nadarajah et al. [29], we can write the density function of Xi:n as
fi:n(x)=r,k=0br,kπr+k+1(x),
where br,k=n!(r+1)(i-1)!υr+1(r+k+1)j=0n-i(-1)jfj+i-1,k(n-i-j)!j!, υr+1 is given in Section 3 and the quantities fj+i−1,k can be determined with fj+i-1,0=υ0j+i-1 and recursively for k ≥ 1
fj+i-1,k=(kυ0)-1m=1k[m(j+1)-k]υmfj+i-1,k-m.
Equation (5.4) reveals that the pdf of the ELOLL-G order statistics is a linear combination of exp-G density functions.

6. Characterizations

This section deals with various characterizations of ELOLL distribution. These characterizations are presented in two directions: (i) based on truncated moments and (ii) in terms of the hazard function. It should be noted that characterization (i) can be employed also when the cdf does not have a closed form. We present our characterizations (i) and (ii) in two subsections.

6.1. Characterizations based on truncated moments

Our first characterization employs a theorem due to Glanzel [16], see Theorem 1 of Appendix B. The result, however, holds also when the interval H is not closed since the condition of Theorem 1 is on the interior of H.

Proposition 1.6

Let X : Ω → ℝ be a continuous random variable and let q1(x)=[G(x)α+G¯(x)α]-2{1-βlog[G¯(x)αG(x)α+G¯(x)α]}-1 and q2(x)=q1(x)[G¯(x)αG(x)α+G¯(x)α] for x ∈ ℝ. The random variable X belongs to the family (2.2) if and only if the function η defined in Theorem 1 has the form

η(x)=θ(θ+1)[G¯(x)αG(x)α+G¯(x)α],x.
Proof. Let X be a random variable with pdf (2.2) , then
(1-F(x))E[q1(x)|Xx]=θ(θ+β)[G¯(x)αG(x)α+G¯(x)α]θ,x,
and
(1-F(x))E[q2(x)|Xx]=θ2(θ+1)(θ+β)[G¯(x)αG(x)α+G¯(x)α]θ+1,x,
and finally
η(x)q1(x)-q2(x)=-1(θ+1)q1(x)[G¯(x)αG(x)α+G¯(x)α]<0,forx.
Conversely, if η is given as above, then
s(x)=η(x)q1(x)η(x)q1(x)-q2(x)=αθg(x)G(x)α-1G¯(x)[G(x)α+G¯(x)α],x,
and hence
s(x)=log{[G¯(x)αG(x)α+G¯(x)α]θ},x.
Now, in view of Theorem 1, X has density (2.2) .

Corollary 1.6.

Let X : Ω → ℝ be a continuous random variable and let q1 (x) be as in Proposition 1. 6. The pdf of X is (2.2) pdf if and only if there exist functions q2 and η defined in Theorem 1 satisfying the differential equation

η(x)q1(x)η(x)q1(x)-q2(x)=αθg(x)G(x)α-1G¯(x)[G(x)α+G¯(x)α],x.
The general solution of the above differential equation is
η(x)={1+G(x)αG¯(x)α}[αθg(x)G(x)α-1G¯(x)α-1[G(x)α+G¯(x)α]2(q1(x))-1q2(x)dx+D],
where D is a constant. Note that a set of functions satisfying the above differential equation is given in Proposition 1. 6. with D = 0. However, it should be also mentioned that there are other triplets (q1, q2, η) satisfying the conditions of Theorem 1.

6.2. Characterization in terms of the hazard function

It is known that the hazard function, hF, of a twice differentiable distribution function, F, satisfies the first order differential equation

f(x)f(x)=hF(x)hF(x)-hF(x).
For many univariate continuous distributions, this is the only characterization available in terms of the hazard function. The following characterization establishes a non-trivial characterization of ELOLL distribution for β = 0, in terms of the hazard function which is not of the above trivial form.

Proposition 2.6

Let X : Ω → ℝ be a continuous random variable. Then, for β = 0, X has pdf (2.2) if and only if its hazard function hF (x) satisfies the differential equation

hF(x)-g(x)g(x)hF(x)=αθg(x)ddx{G(x)α-1G¯(x)[G(x)α+G¯(x)α]},x.
Proof. If X has pdf (2.2), then clearly the above differential equation holds. Now, if the differential equation holds, then
ddx{(g(x))-1hF(x)}=αθddx{G(x)α-1G¯(x)[G(x)α+G¯(x)α]}
or
hF(x)=αθg(x){G(x)α-1G¯(x)[G(x)α+G¯(x)α]},x,
which is the hazard function of (2.2).

7. Maximum Likelihood Estimation (MLE) and Inference

Several approaches for parameter estimation were proposed in the literature but the maximum likelihood method is the most commonly employed. The MLEs enjoy desirable properties and can be used for constructing confidence intervals and also for test statistics. The normal approximation for these estimators in large samples can be easily handled either analytically or numerically. Here, we consider the estimation of the unknown parameters of the new family from complete samples only by maximum likelihood. Let x1,...,xn be a random sample from ELOLL-G model with a (q+3)×1 parameter vector Ξ =(α, β, θ, ξ) , where ξ is a q×1 baseline parameter vector. The log-likelihood function for Ξ can be expressed as

n(Ξ)=nlog(α)+2nlog(θ)-nlog(θ+β)+i=1nlogg(xi,ξ)+(αθ-1)i=1nlogG¯(xi,ξ)+(α-1)i=1nlogG(xi,ξ)+(1-θ)i=1nlog(ti)+i=1nlog{1-βlog[G¯(xi,ξ)αti]},
where ti=G(xi,ξ)α+G¯(xi,ξ)α. The log-likelihood function can be maximized by solving the following nonlinear normal equations which are available if needed. To solve these equations, it is more convenient to use nonlinear optimization methods such as the quasi-Newton algorithm to numerically maximize (Ξ).

The likelihood ratio (LR) statistic can be used for comparing the some sub-models of ELOLLG model. For example, the LR statistic can be used to discriminate between the ELLOL-Lomax and L-Lomax since they are nested models, which is equivalently to test H0 : α = β = 1. The LR statistic reduces to w=2[(α^,β^,σ^,β^)-(1,1,σ˜,β˜)], where (α^,β^,σ^,β^) are the unrestricted MLEs and (1,1,σ˜,β˜) are the restricted estimates under H0. The statistic w is asymptotically (as n → ∞) distributed as χk2, where k is difference of two parameter vectors of nested models. For example, k = 2 for above hypothesis test.

8. A simulation study

In this subsection, we perform the simulation study using the exponential Lindley odd log-logistic exponential distribution (ELOLL-E) which is the generalization of the exponential distribution with cdf F (x; λ) = 1 − exp(−λx), x > 0, λ > 0. To see the performance of MLE’s of this distributions, we generate 1,000 samples of sizes 20, 50 and 100 from ELOLL-E using inverse of the its cdf for θ = 2. The results of the simulation are reported in Table (2). From this Table, we observe that the estimates approach true values as the sample size increases whereas the standard devations of the estimates decrease.

Parameters n = 20 n = 50 n = 100

α, β, λ α^,β^,λ^ α^,β^,λ^ α^,β^,λ^
1,1,1 1.0899,1.1654,1.0906 1.0170,1.0452,1.0189 1.0111,1.0031,0.9847
(0.2581),(0.9818),(0.3488) (0.1379),(0.8293),(0.2274) (0.0848),(0.7345),(0.1670)
2,2,2 2.1705,2.3459,1.9854 2.0673,1.9774,1.9883 2.0361,1.9851,1.9916
(0.4130),(1.9485),(0.2642) (0.2550),(0.7811),(0.1546) (0.1795),(0.5796),(0.1736)
2.1.0.5 2.1053,1.0154,0.5046 2.0586,1.0128,0.4996 2.0117,1.0085,0.5009
(0.3937),(0.3690),(0.0603) (0.2405),(0.3371),(0.0424) (0.1653),(0.2717),(0.0286)
1,0.5,2 1.0147,0.7305,2.1798 1.0062,0.6800,2.0887 0.9945,0.6546,2.0635
(0.2305),(0.8415),(0.7783) (0.1214),(0.6699),(0.4054) (0.0912),(0.5922),(0.4099)
1,2,0.5 1.0834,1.8239,0.5047 1.0284,1.9398,0.5030 1.0210,1.9620,0.5017
(0.2335),(0.6285),(0.1195) (0.1259),(0.4712),(0.0764) (0.0872),(0.4550),(0.0530)
2,0.5,0,5 2.0795,0.5410,0.5105 2.0380,0.5515,0.5035 2.0156,0.5280,0.5033
(0.3949),(0.2886),(0.0661) (0.2302),(0.6220),(0.0445) (0.1683),(0.5280),(0.0295)
0.75,5,2 0.8201,4.8903,2.1585 0.7895,4.8956,1.9550 0.7619,4.9821,1.9942
(0.1484),(0.7185),(0.5310) (0.1029),(0.3244),(0.3808) (0.0659),(0.2595),(0.2433)
10,10,10 10.679,10.0500,10.0436 10.1903,10.0244,10.0041 10.1852,9.9983,9.9934
(2.0187),(2.1032),(0.2139) (1.1108),(1.5571),(0.1350) (0.9405),(2.2128),(0.1038)
Table 2:

Emprical means and standard deviations (in paranthesis) for selected parameters values of ELOLL-E distribution

9. Data analysis

In this section, we introduce three applications using well-known data sets to show the flexibility and applicability of the proposed models over other models. First, we describe the data sets and we fit some distributions to these data sets using MLE method. Then we compare proposed distributions with several member of distribution families. The model selection is applied using the estimated log-likelihood (^), Kolmogorov-Smirnov (K-S) statistics, Akaike information criterion (AIC), Consistent Akaike information criteria (CAIC), Bayesian information criterion (BIC), and Hannan-Quinn information criterion (HQIC). The AIC, CAIC, BIC and HQIC are by given by AIC=-2^+2p, CAIC=-2^+2pn(n-k-1)-1, BIC=-2^+plogn and HQIC=-2^+plog(logn), where p is the number of the estimated model parameters and n is sample size. When searching the best fit among others to data, the distribution with the smallest AIC, CAIC, BIC, HQIC and K-S values and the biggest log-likelihood and p values of the K-S statistics is chosen. All calculations are obtained by maxLik routine in R programme.

First data set studied by Murthy et al. [27], which represent failure times for a particular windshield device. This data set has been analyzed by Cordeiro et al. [14]. Using this data set, we fit the ELOLL-N, Lindley-normal (L-N) (Çakmakyapan and Ozel, [9]), exponential-normal (E-N) or Lehmann type II exponentiated-normal (Alzaatreh et al., [5]; Cordeiro et al. [15]), McDonaldnormal (McN) (Alexander et al. [3]), normal-normal{exponential} (NNE) (Alzaatreh et al. [6]), normal-Cauchy{log-logistic} (NCLL) (Alzaatreh, et al. [7]), logistic-normal (LN) (Tahir et al. [32]), generelized Kumaraswamy-normal (GKw-N) (Cordeiro et al. [11]) and generalized odd log-logistic normal (GOLLN) (Cordeiro, et al. [10]) distributions models. The results of this application are listed in Table 3.

Model α^,β^,θ^,μ^,σ^ -^ AIC,CAIC,BIC,HQIC K-S (p-v.)
ELOLL-N 0.1679, 6.5946, 3.0515, 3.1785, 0.4070 124.3870 258.7740, 259.5433, 270.9281, 263.6599 0.0463
(0.0849), (9.5475), (0.6488), (0.2077), (0.1255) (0.9937)
McN 125.6276, 6.9936, 0.0299, 4.7966, 2.4578 128.2184 266.4367, 267.2059, 278.5908, 271.3225 0.0805
(0.2334), (0.3941), (0.0131), (0.9119), (0.4754) (0.6478)
NNE 9.0338, 0.7306,−19.3314, 5.9635 128.4202 264.8404, 265.3467, 274.5637, 268.7491 0.0991
(0.4896), (0.0873), –, (0.4711), (0.3606) (0.3816)
NCLL 14.6194, 0.7789,−19.0441, 4.4038 128.1291 264.2582, 264.7645 ,273.9814, 268.1669 0.0921
(3.8058), (0.1874), –, (2.6534), (0.9849) (0.4742)
LN –, –, 6.1695, 1.1181, 4.0953 130.0301 266.0602, 266.3602, 273.3526, 268.9917 0.0805
–, –,(4.5532), (1.0089), (3.0241) (0.6482)
GKwN 90.5274, 4.6255, 0.6399,−6.3998, 4.0251 127.9231 265.8463, 266.6155, 278.0003, 270.7321 0.0971
(13.5280), (5.0395), (0.3384), (1.7059), (1.2968) (0.4067)
GOLLN 0.4557, 1.1863, –, 2.5139, 0.6358 127.0450 262.0900, 262.5963, 271.8133, 265.9987 0.0901
(0.2353), (1.2905), –, (0.7508), (0.3306) (0.5036)
L-N –, –, 0.1446, 0.258, 0.5530 127.8708 261.7416, 262.0416, 269.0341, 264.6731 0.0775
–, –, (0.0112), (1.3e-5), (6.2e-6) (0.6946)
E-N –, –, 0.1382, 1.0954, 0.5583 128.0645 262.1289, 262.4289, 269.4214, 265.0604 0.0618
–, –, (0.0151), (0.0006), (0.0002) (0.8908)
Table 3:

MLEs of the model parameters for the windshield data, the corresponding standard errors (given in parentheses below estimated parameters) and the AIC, CAIC, BIC, HQIC and K-S values

As second data analysis, we analyze the data set studied by Abouammoh et al. [1], which represent the lifetime in days of 40 patients suffering from leukemia from one of the Ministry of Health Hospitals in Saudi Arabia. By using this data set we compare the ELOLL-W distribution with logistic-Weibull (LW) (Tahir et al. [32]), odd-Burr-Weibull (OBW) (Alizadeh et al. [4]), Lindley-Weibull (L-W) (Çakmakyapan and Ozel [9]) and exponential-Weibull (E-W) or ordinary Weibull distributions. The results of this application are listed in Table 4.

Model α^,β^,θ^,λ^,γ^ -^ AIC,CAIC,BIC,HQIC K-S(p-v.)
ELOLL-W 0.2389, 44 1.841 6, 1. 7418 , 0.00 09, 4.5233 299.1043 608.2087, 609.9734, 616.6531, 611.2619 0.0651
(0.0567), (4.1943), (0.4284), (0.0001), (0.4813) (0.9958)
LW –,–, 34.7613, 0.0009, 0.0916 311.4651 628.9302, 629.5969, 633.9968, 630.7621 0.1618
–,–, (3.8176), (0.0000), (0.0016) (0.2459)
OBW 3.0365, –, 28.9395, 0.00015, 0.7735 303.8669 615.7339, 616.8768, 622.4894, 618.1765 0.1163
(1.3087), –, (1.1964), (0.0000), (0.3973) (0.6520)
L-W 1, 1, 0.8129, 0.0011, 2.3265 303.6019 613.2038, 613.8704, 618.2704, 615.0357 0.1075
–, –, (0.9133), (0.0006), (0.4141) (0.7436)
E-W 1, 0, 336.0406, 0.00008, 2.5775 304.4153 614.8305, 615.4972, 619.8972, 616.6625 0.1237
–, –, (4.2136), (0.00002), (0.4103) (0.5728)
Table 4:

MLEs of the model parameters for the lukemia data, the corresponding standard errors (given in parentheses below estimated parameters) and the AIC, CAIC, BIC, HQIC and K-S values

The third real data consists of the number of successive failure for the air conditioning system reported of each member in a fleet of 13 Boeing 720 jet airplanes. The pooled data with 213 observations was considered by Proschan [30] and Kuş [23]. For this data, we use the ELOLL-Lx distribution and compare it Zografos-Balakrishan odd-loglogistic-Lomax (ZBOLL-Lx) (Cordeiro et al. [13]), Kumaraswamy Lomax (Kw-Lx) (Lemonte and Cordeiro [24]), beta-Lomax (B-Lx) (Lemonte and Cordeiro [24]) and its sub-models. The results of the application are in Table 5.

Model α^,β^,θ^,k^,λ^ -^ AIC,CAIC,BIC,HQIC K-S (p-v.)
ELOLL-Lx 5.1293, 42.3803, 193.8547, 0.0461, 0.0879 1172.9880 2355.9770, 2356.2670, 2372.7830, 2362.7690 0.0372
(1.4911), (2.1788), (1.1108), (0.0032), (0.1794) (0.9296)
Kw-Lx 1.1322, 234.9295, -, 0.0208, 172.2718 1174.9830 2357.9660, 2358.1590, 2371.4110, 2363.4000 0.0439
(0.0622), (0.3286), -, (0.0053), (2.1826) (0.8043)
BLx 0.9395, 13.5670, - (0.0053), (2.1826) 1176.7130 2361.4260, 2363.7160, 2380.2330, 2370.2180 0.0583
(0.0794),(0.6603),-, (0.2486),(2.1986) (0.4642)
ZBOLL-Lx 14.9281, 0.4497,-, 0.0605, 0.0013 1177.1000 2362.1990, 2362.3920, 2375.6440, 2367.6330 0.0489
(1.8577), (0.0878),-, (0.0028), (0.0006) (0.6878)
L-Lx 1, 1, 441.4941, 0.0130, 441.9408 1175.3800 2356.8750, 2356.8750, 2366.8440, 2360.8350 0.0389
1, 1, (2.1022), (0.0009), (0.8563) (0.9040)
E-Lx 1, 0, 1346.2550, 0.0045, 469.5059 1175.3850 2356.7500, 2356.8850, 2366.8540, 2360.8450 0.0390
1, 0, (2.3463), (0.0003), (1.0493) (0.9021)
OBLx 1.0194, -, 1251.0490, 0.0052, 444.8587 1175.3160 2358.632, 2358.8250, 2372.0780, 2364.0660 0.0403
(0.0544),-, (2.1920), (0.0018), (0.6342) (0.8796)
EL-Lx 1, 78.9168, 1361.9960, 0.0043, 433.0816 1175.3780 2358.7560, 2358.9480, 2372.2010, 2364.1900 0.0390
1, (2.0991), (0.7953), (0.0003), (0.7930) (0.9036)
LOLL-Lx 1.1071, 1,396.4973, 0.0131, 201.9921 1174.9910 2357.9810, 2358.1740, 2371.4260, 2363.4150 0.0433
(0.0587), -, (2.1042), (0.0035), (2.1207) (0.8181)
Lx 1, 0, 1, 22085.62, 1930929.40 1178.3410 2360.6830, 2360.7400, 2367.4060, 2363.4000 0.2346
-, -, -, (23.9112), (00000) (00000)
Table 5:

MLEs of the model parameters for the Proschan data, the corresponding standard errors (given in parentheses below estimated parameters) and the AIC, CAIC, BIC, HQIC and K-S values

From Tables 35, we see that the ELOLL-N, ELOLL-W and ELOLL-Lx models have the lowest AIC, CAIC, BIC, HQIC and K-S values and has the biggest estimated log-likelihood and p-value of the K-S statistics among all the fitted models. So they could be chosen as the best models under these criteria.

The histograms of three data sets and the estimated pdfs and cdfs of the application models are displayed in Figures 4. From the Figures 4, we show that the ELOLL-G models provide the good fit to these data sets as compared to other models.

Fig. 4:

The fitted pdfs and the fitted cdfs for three data sets

A comparison of the proposed distributions with some of their sub-models using LR statistics is performed in Table 6. From Table 6, we can conclude that the ELOLL-N and ELOLL-W models yield a better fit to these data than the other two and three sub-models respectively. Also, there is no difference among the fits to the current data using the ELLOL-Lx, L-Lx and E-Lx models. In addition, these models provide a better representation of the data than the other sub-models based on the LR test at the 5% significance level.

Model Hypothesis Test statistics p-value

ELLOL-N vs L-N H0:α=β=1, H1: H0 false 6.9668 0.0307
ELLOL-N vs E-N H0:α=1, β=0, H1: H0 false 7.3550 0.0253

Model

ELLOL-W vs L-W H0:α=β=1, H1: H0 false 8.9952 0.01111
ELLOL-W vs E-W H0:α=1, β=0, H1: H0 false 10.622 0.00494
ELLOL-W vs H0: β=0, H1: H0 false 9.5252 0.00202

Model

ELLOL-Lx vs L-Lx H0: α=β=1, H1: H0 false 4.7840 0.09144
ELLOL-Lx vs E-Lx H0: α=1, β=0, H1: H0 false 4.7940 0.09099
ELLOL-Lx vs OBLx H0: β=0, H1: H0 false 4.6560 0.03094
ELLOL-Lx vs EL-Lx H0: α=1, H1: H0 false 4.7800 0.02879
ELLOL-Lx vs LOLL-Lx H0: β=1, H1: H0 false 4.0060 0.04533
ELLOL-Lx vs Lx H0:α=θ=1, β=0, H1: H0 false 10.7060 0.013426
Table 6:

LR statistics for three data sets

10. Conclusions

A new family of distributions called the exponential Lindley odd log-logistic G family is introduced and studied. The new family generalizes the Lindley-G family [9], Lehmann Type II-G family [19] and odd Burr-G family [4] as well as it introduces new distribution families such as exponential Lindley-G family and Lindley odd log-logistic-G family. We provide some mathematical properties of the new family including ordinary, generating function and order statistics. Characterizations based on truncated moments as well as in terms of the hazard function are presented. The maximum likelihood is used for estimating the model parameters. We give a simulation study for the maximum likelihood estimators by using the special member of new family. Finally, the usefulness of the family is illustrated by means of three real data sets. The new models provide consistently better fits than other competitive models for these data sets.

Appendix A.

A=k=0(-1)k(αθk)dkG(x)k,B=k=0bkG(x)k,C=k=0hkG(x)k,
and
D=k=0l=k(-1)l+k(α(i+1)+jl)(lk)qkG(x)k.
Since [G(x)c+G¯(x)c]=m=0tmG(x)m, where tm=(-1)m[(cm)+w=m(-1)w(cw)(cm)]. Then
[G(x)c+G¯(x)c]a=w=0k=0(-1)k-wk!(kw)(a)wfw(m=0tmG(x)m)w,
where (a)w = a(a − 1) ... (ak + 1) is the descending factorial. For a power series raised to a positive integer w(m=0tmum)w=m=0mw,mum, where the coefficients m w,m (for m = 1.2,...) are easily determined from the recurrence equation mw,m=(mt0)-1d.=1m[d(m+1)-m](td)(mw,m-d), where mw,0=t0w, The coefficient m w,m can be calculated from m w,0,...,m w,m−1 and hence from the quantities t0,...,tm. So, [G(x)α+G¯(x)α]θ=k=0bkG(x)k, where bk=bk(α,θ)=w=0fwmw.k, and [G(x)α+G¯(x)α]θ+i+1=m=0hkG(x)k,where hk=hk(α,θ+i+1)=w=0fwmw.k.

Appendix B.

Theorem 1.

Let (Ω, , P) be a given probability space and let H = [d, e] be an interval for some d < e (d = −∞, e = ∞ might as well be allowed). Let X : Ω → H be a continuous random variable with the distribution function F and let q1 and q2 be two real functions defined on H such that

E[q2(X)|Xx]=E[q1(X)|Xx]η(x),xH,
is defined with some real function η. Assume that q1, q2C1 (H), ηC2 (H) and F is twice continuously differentiable and strictly monotone function on the set H. Finally, assume that the equation ηq1 = q2 has no real solution in the interior of H. Then F is uniquely determined by the functions q1, q2 and η, particularly
F(x)=axC|η(u)η(u)q1(u)-q2(u)|exp(-s(u))du,
where the function s is a solution of the differential equation s=ηq1ηq1-q2 and C is the normalization constant, such that HdF=1.

References

[4]M Alizadeh, GM Cordeiro, ADC Nascimento, MDS Lima, and EMM Ortega, Odd-Burr generalized family of distributions with some applications, Journal of Statistical Computation and Simulation, Vol. 83, 2017, pp. 326-339.
[9]S Çakmakyapan and G Ozel, The Lindley Family of Distributions: Properties and Applications, Hacettepe Journal of Mathematics and Statistics, 2017.
[11]GM Cordeiro, M Alizadeh, RB Silva, and TG Ramires, A new wider family of continuous models: The extended Cordeiro and De Castro family, Hacettepe Journal of Mathematics and Statistics, 2017.
[13]GM Cordeiro, M Alizadeh, EM Ortega, and LHV Serrano, The Zografos-Balakrishnan odd log-logistic family of distributions: Properties and Applications, Hacet. J. Math. Stat, Vol. 45, 2016, pp. 1781-1803.
[17]JU Gleaton and JD Lynch, Properties of generalized log-logistic families of lifetime distributions, Journal of Probability and Statistical Science, Vol. 4, 2006, pp. 51-64.
[20]GG Hamedani, HM Yousof, M Rasekhi, M Alizadeh, and SM Najibi, Type I general exponential class of distributions, International Journal of Applied & Experimental Mathematics, 2017. forthcoming.
[26]AM Mathai and RK Saxena, The H-Function with Applications in Statistics and Other Disciplines, John Wiley and Sons, New York, 1978.
[27]DNP Murthy, M Xie, and R Jiang, Weibull Models, 1st ed., John Wiley & Sons, Hoboken, NJ, 2004.
[31]HM Srivastava, KC Gupta, and SP Goyal, The H-Functions of One and Two Variables with Applications, South Asian Publishers, New Delhi, 1982.
[35]HM Yousof, AZ Afify, GG Hamedani, and G Aryal, the Burr X generator of distributions for lifetime data, Journal of Statistical Theory and Applications, Vol. 16, 2016, pp. 1-19.
[36]HM Yousof, M Rasekhi, AZ Afify, M Alizadeh, I Ghosh, and GG Hamedani, The beta Weibull-G family of distributions: theory, characterizations and applications, Pakistan Journal of Statistics, 2017. forthcoming
Journal
Journal of Statistical Theory and Applications
Volume-Issue
17 - 3
Pages
554 - 571
Publication Date
2018/09/30
ISSN (Online)
2214-1766
ISSN (Print)
1538-7887
DOI
10.2991/jsta.2018.17.3.10How to use a DOI?
Copyright
© 2018, the Authors. Published by Atlantis Press.
Open Access
This is an open access article under the CC BY-NC license (http://creativecommons.org/licences/by-nc/4.0/).

Cite this article

TY  - JOUR
AU  - Mustafa Ç. Korkmaz
AU  - Haitham M. Yousof
AU  - G. G. Hamedani
PY  - 2018
DA  - 2018/09/30
TI  - The Exponential Lindley Odd Log-Logistic-G Family: Properties, Characterizations and Applications
JO  - Journal of Statistical Theory and Applications
SP  - 554
EP  - 571
VL  - 17
IS  - 3
SN  - 2214-1766
UR  - https://doi.org/10.2991/jsta.2018.17.3.10
DO  - 10.2991/jsta.2018.17.3.10
ID  - Korkmaz2018
ER  -