Submit manuscript...
eISSN: 2378-315X

Biometrics & Biostatistics International Journal

Research Article Volume 11 Issue 1

Reliability estimation of type-II generalized loglogistic distribution

SVSVSV Prasad,1 Gadde Srinivasa Rao,2 K. Rosaiah1

1Department of Statistics, Acharya Nagarjuna University, India
2Department of Mathematics and Statistics, The University of Dodoma, Tanzania

Correspondence: Gadde Srinivasa Rao, Department of Mathematics and Statistics, The University of Dodoma, Dodoma, P.O. Box: 259, Tanzania

Received: December 20, 2021 | Published: February 28, 2022

Citation: Prasad SVSVSV, Rao GS, Rosaiah K. Reliability estimation of type-II generalized log-logistic distribution. Biom Biostat Int J. 2022;11(1):36-46. DOI: 10.15406/bbij.2022.11.00352

Download PDF

Abstract

In this article, a lifetime distribution named as Type-II generalized log-logistic distribution (TGLLD) is considered and its failure rate of products with different shape parameters used to find out ageing criteria. An attempt has been made to derive the statistical and reliability properties of TGLLD. Parameters are evaluated using maximum likelihood estimation and obtained the reliability of the distribution. A simulation study also conducted to know the performance of the estimators. The estimates obtained are validated with the use of live data.

Keywords: type-ii generalized log-logistic distribution, reliability function, hazard rate, reverse hazard rate, moments, maximum likelihood estimation

MATHEMATICS SUBJECT CLASSIFICATION: 62F10; 62F15; 62N05

Introduction

Over a period of time, statistical literature witnessed the origin of many continuous univariate distributions. However, in the present era, these distributions are extended by introducing the additional parameters in order to cater the requirements from different areas such as lifetime analysis, finance, engineering industries, insurance etc. The present distribution dealt in this article is one such distribution introduced by Rosaiah et al.1 When a distribution is introduced, one may keen to know the behavior for its characterization. The same can be achieved by finding its statistical properties viz., mean, median, mode, variance, quantiles, moments, cumulants, order statistics, ML estimates, confidence intervals etc. Distinguished authors have made their efforts in estimating such properties for different distributions viz. Balakrishnan2,3 for half logistic and generalized logistic, Mudholkar and Srivastava4 Mudholkar5 for exponentiated Weibull, Gupta et al.6 for log-logistic, Nadarajah7 for exponentiated Gumbel, Nadarajah and Gupta8 for exponentiated gamma, Abouammoh and Alshingiti9 for inverted exponential distribution, Rosaiah et. al.10 for odds exponential log logistic Distribution and many more.

Log-logistic distribution (LLD) has proven its importance in quality control, mainly in analyzing the lifetime data. Many authors have made their contribution in developing the various features of this distribution by creating some extensions to the original distribution, Type-II generalized log-logistic distribution (TGLLD) is one such distribution. In this article an effort to derive mathematical properties of TGLLD. The rest of the article is organized as follows. In Section 2, the cumulative distribution function, probability density function, reliability function and hazard function of TGLLD are given. Also, the key properties, moments of TGLLD and ith order statistic are obtained. In Section 3, ML estimators, asymptotic confidence intervals are derived. Fitting reliability data, computation of ML estimates, statistics such as -2logL, AIC and BIC are presented in section 4. Lastly in Section 5, the concluding remarks are given.

Type-II generalized log logistic distribution

Log-logistic distribution (LLD) has proven its importance in quality control. Different authors developed properties and types of acceptance sampling plans for LLD viz., Ashkar and Mahdi.11 The cumulative distribution function (CDF) of the log-logistic distribution (LLD) is

F(t;σ,θ)=(t/σ)λ[1+(t/σ)λ];t>0,σ>0,λ>1F(t;σ,θ)=(t/σ)λ[1+(t/σ)λ];t>0,σ>0,λ>1                                                  (1)

Since the practical pertinence of generalized log-logistic distribution (GLLD) in diverse sectors, various authors have paid their attention in developing some extensions for effective and wide use of log-logistic distribution. One such extension to this distribution named as Type-II generalized log-logistic distribution (TGLLD) introduced by Rosaiah et al.,1 its cumulative distribution function (cdf) is

F(t;σ,θ,λ)=1[1+(t/σ)λ]θ;t>0,σ>0,θ>0,λ>1  (2)

It may be noted that the distribution given in (2) is defined through the reliability oriented generalization of log-logistic distribution. In short, we call this as the Type-II generalized log-logistic distribution [Type-I generalized (exponentiated) log-logistic distribution is dealt by Rosaiah et al.12 The corresponding probability density function (PDF) is given by

f(t;σ,θ,λ)=λθσ(t/σ)λ1[1+(t/σ)λ]θ+1;t>0,σ>0,θ>0,λ>1                                                                      (3)           

where σ is the scale parameter, λ and θ are shape parameters.

Rao et al.13,14 developed the reliability test plans for this distribution. The reliability function and hazard (failure rate) function of type-II generalized log-logistic distribution are respectively given by

R(t)=[1+(t/σ)λ]θ                                                                                                                                               (4)

h(t)=λθ(t/σ)λ1[1+(t/σ)λ];t>0,σ>0,θ>0,λ>1                                                                                            (5)      

The three-parameter TGLLD will be denoted by TGLLD (σ,θ,λ) . If θ=1 , then Eq. (3) becomes log-logistic distribution and if λ =1 then TGLLD becomes reduced to Pareto type-II distribution. Figures 1-4 depicted that the PDF, CDF, reliability function and hazard function curves of TGLLD for various parametric combinations.

Figure 1 The probability density function of TGLLD for different λ and θ  at σ .

Figure 2 The cumulative density function of TGLLD for different λ and θ  at σ .

Figure 3 The reliability function of TGLLD for different λ and θ  at σ .

Figure 4 The hazard function of TGLLD for different λ and θ  at σ . (a) Upside-down bathtub (b) Increasing (c) Decreasing.

Properties of the TGLLD

Limits of the distribution function

F(t;σ,θ,λ)=1[1+(t/σ)λ]θ;t>0,σ,θ>0,λ>1

limt0F(t;σ,θ,λ)=limt0{1[1+(t/σ)λ]θ}=0

limtF(t;σ,θ,λ)=limt{1[1+(t/σ)λ]θ}=1

Reverse hazard function of a non-negative random variable is r(t)=f(t)F(t)=λθσ(t/σ)λ1[1+(t/σ)λ]θ[[1+(t/σ)λ]θ1] (6)

The odd function of a random variable can be derived from

O(t)=F(t)h(t)=λθ(t/σ)λ1[1+(t/σ)λ]θ1[1+(t/σ)λ]θ1                                                                                        (7)

Mean of TGLLD is derived as

μ=0tλθσ(t/σ)λ1[1+(t/σ)λ]θ+1dt=σλΓ(1λ)Γ(θ1λ)Γ(θ)            (8)

Percentile and median

The 100pth percentile of a random variable T is denoted by tp and is defined as

tp=inf{t:F(t)p}, where F(t;σ,θ,λ)=1[1+(t/σ)λ]θ

If t;tp  is unique for each p(0,1),F1(p)  is an inverse function, thentp=F1(p)

100pth Percentile, p=F(t)=1[1+(t/σ)λ]θ

tp=σ[(1p)1/θ1]1/λ                                                                                                  (9)

Quantile function Q(p)=σ[(1p)1/θ1]1/λ

Median (M) (50th percentiles) is

M=σ[21/θ1]1/λ                                                                                        (10)

Mode of TGLLD is obtained as the value of x for which logf(t)t=0 and 2logf(t)t20 .

logf(t)==logλ+logθlogσ+(λ1)log(t/σ)(θ+1)log[1+(t/σ)λ]

logf(t)t=(λ1)(1/t)(θ+1)1[1+(t/σ)λ]λ(t/σ)λ1

2logf(t)t2={(λ1)t2+(θ+1)[λ2σ(tσ)λ1λ(t/σ)λ(1+(t/σ)λ)2]}0

Hence mode is the solution of the non-linear equation

(λ1)tλ(θ+1)(t/σ)λ1[1+(t/σ)λ]=0

Moments of TGLLD

Moments are the useful tools which can be used to derive the key features of the distribution. Here we tried to derive rth moment of the random variable T, where TTGLLD(θ,λ,σ) .

The rth moment of T is denoted by μ'r  and is defined as

μ'r=E(tr)=0trλθσ(t/σ)λ1[1+(t/σ)λ]θ+1dt=rλσrΓ(rλ)Γ(θrλ)Γ(θ)                                                                         (11)

Variance = μ2=σ2λΓ(θ)(2Γ(2λ)Γ(θ2λ)Γ(θ)1λ(Γ(1λ)Γ(θ1λ))2Γ(θ))

μ3=σ3λΓ(θ)(3Γ(3λ)Γ(θ3λ)6λΓ(1λ)Γ(2λ)Γ(θ1λ)Γ(θ2λ)Γ(θ)+2λ2(Γ(1λ)Γ(θ1λ))3(Γ(θ))2)

μ4=σ4λΓ(θ)[4Γ(4λ)Γ(θ4λ)12λΓ(1λ)Γ(3λ)Γ(θ1λ)Γ(θ3λ)Γ(θ)+12λ2Γ(2λ)Γ(θ2λ)[Γ(1λ)Γ(θ1λ)]2(Γ(θ))23λ3[Γ(1λ)Γ(θ1λ)]4(Γ(θ))3]

Mean, median, skewness and kurtosis of TGLLD for various combinations of λ , θ  and σ  are given in Table 1.

λ

θ

σ

Mean

Median

Skewness

Kurtosis

1.5

1.5

1

1.1498

0.7014

4.0543

3.0912

2.5

2.5

1

0.6985

0.6336

2.5365

9.9632

3

3

1

0.6718

0.6382

0.8453

5.1318

3.5

3.5

1

0.6657

0.648

0.3195

3.8556

4

4

1

0.6682

0.6595

0.112

3.3635

4.5

4.5

1

0.6744

0.6714

0.0283

3.1535

5

5

1

0.6824

0.6831

0.0016

3.07

5.5

5.5

1

0.6911

0.6942

0.0039

3.0504

6

6

1

0.6999

0.7047

0.0215

3.0653

2

3

2

1.1781

1.0196

3.6429

12.4635

3

2

3

2.4184

2.2363

2.5253

10.8095

2

6

2

0.7731

0.6999

1.1978

5.1176

6

2

3

2.618

2.5902

0.1888

4.1057

Table 1 Mean, Median, Skewness and Kurtosis of TGLLD for various combinations of λ, θ andσ

Moment Generating function (MGF) is given by

Mt(z)=E(etz)=r=0zrr!σr1(rλ)!(θrλ1)!(θ1)!                                                      (12)

Characteristic Function is given by                         

ϕt(z)=E(eizt)=r=0(iz)rr!σr1(rλ)!(θrλ1)!(θ1)!                                         (13)

Cumulative Generating function is defined by        

Kt(z)=ln(Mt(z))=Kt(z)=ln(r=0zrr!σr1(rλ)!(θrλ1)!(θ1)!)      (14)

Order statistics of TGLLD

Let T1:nT2:n..........Tn:n denotes the order statistics obtained from a random sample of size n drawn from TGLLD (θ,λ,σ). The probability density function of ith order statistic is given by                                 

fi:n(t;θ,λ,σ)=1β(i,ni+1)(F(t))i1(1F(t))nif(t)             (15)

Since 0<F(t;θ,λ,σ)<1 for t>0 , then using binomial expansion

(F(t))i1=i1j=0(i1j)(1)j(1F(t))j          (16)

Then fi:n(t;θ,λ,σ)=1β(i,ni+1)f(t)i1j=0(i1j)(1)j(1F(t))n+ji                     (17)

(1F(t))n+ji=[1+(t/σ)λ]θ(n+ji)

Now take f(t)[1f(t)]n+ji=λθ(t/σ)λ1σ[1+(t/σ)λ]θ+1[1+(t/σ)λ]θ(n+ji)                                                

=λθσ(t/σ)λ1[1+(t/σ)λ]θ(n+ji+1)1

Hence fi:n(t;λ,θ,σ)=i1j=0(1)jn!j!(ni)!(ij1)!f(t;λ,θ,σ,(n+ji+1))       (18)

The distribution function of ith order statistic T(i) is

Fi:n(t;σ,θ,λ)=nj=i(ni)Fj(t)[1F(t)]nj , using (16), it can be expressed as

Fi:n(t;λ,θ,σ)=nj=ijk=0(ni)(jk)(1)k(1+(t/σ)λ)θ(n+kj)      (19)

Distribution function of first order statistic T(1) is

F1(t)=1(1F(t))n=1(1+(t/σ)λ)nθ  (20)

Distribution function of nth order statistic T(n) is                              

Fn(t)=(F(t))n=(1(1+(t/σ)λ)θ)n                                                                                                   (21)

Parameter estimation and inference

For estimating the parameters of TGLLD(σ,θ,λ) , we considered two known methods viz., maximum likelihood method of estimation and least square method. It is observed that the estimates obtained from both methods for the unknown parameters cannot be expressed in closed form and hence the estimates are obtained using simulation study.

Maximum likelihood estimators (MLEs)

Let t1,t2.....tn be a random sample of size n drawn from TGLLD (T;θ,λ,σ) , then likelihood function L of the sample is

L=ni=1f(ti;θ,λ,σ)=ni=1λθσ(ti/σ)λ1[1+(ti/σ)λ]θ+1

The log-likelihood function is

logL=nlogλ+nlogθnlogσ+(λ1)ni=1log(ti/σ)(θ+1)ni=1log[1+(ti/σ)λ]    (22)

The MLE’s of θ,λ  and σ are obtained as

logLσ=0nλσ+λ(θ+1)σni=1(ti/σ)λ[1+(ti/σ)λ]=0                           (23)       

logLλ=0nλ+ni=1log(ti/σ)(θ+1)ni=1(ti/σ)λlog(ti/σ)[1+(ti/σ)λ]=0                         (24)

logLθ=nθni=1log[1+(ti/σ)λ]=0ˆθ=nni=1log[1+(ti/σ)λ]                     (25)

Using Eq. (25) in Eqs. (23) and (24) we get two equations in terms of σandλ , these equations cannot be solved analytically, so they need to be solved numerically. Iterative techniques can be applied for obtaining the estimators of the parameters. Let ˆσ,ˆλandˆθ  are ML estimates of the parameters σ,λandθ respectively. Using invariance property of the MLE, the MLE of reliability function can be obtained by

ˆR(t)=[1+(t/ˆσ)ˆλ]ˆθ       (26)

Asymptotic confidence interval

Here, an attempt has been made to derive the asymptotic confidence intervals of the unknown parameters θ,λ and σ . Using large sample approach and assume that the MLE’s of (ˆθ,ˆλandˆσ)  are approximately multivariate normal with mean (θ,λ,σ) and Variance-covariance matrix I1 , where I1 is observed information matrix which is defined as

I1=[2logLθ22logLλθ2logLσθ2logLθλ2logLλ22logLσλ2logLθσ2logLλσ2logLσ2] =[var(ˆθ)cov(ˆλ,ˆθ)cov(ˆσ,ˆθ)cov(ˆθ,ˆλ)var(ˆλ)cov(ˆσ,ˆλ)cov(ˆθ,ˆσ)cov(ˆλ,ˆσ)var(ˆσ)]    (27)

Now, the second order partial derivatives of the parameters given in I1 are

2logLθ2=θ[logLθ]=nθ2 (28)

2logLλθ=λ[logLθ]=ni=1(ti/σ)λlog(ti/σ)[1+(ti/σ)λ] (29)

2logLσθ=σ[logLθ]=ni=1λσ(ti/σ)λ[1+(ti/σ)λ] (30)

2logLθλ=θ[logLλ]=ni=1(ti/σ)λlog(ti/σ)[1+(ti/σ)λ] (31)

2logLλ2=λ[logLλ]=[nλ2(θ+1)ni=1{(ti/σ)λ[1(ti/σ)λ][log(ti/σ)]2[1+(ti/σ)λ]2}] (32)

2logLσλ=σ[logLλ]nσ(θ+1)ni=1{tiλσ(λ+1)[λlog(ti/σ)1ti1ti(ti/σ)λ][1+(ti/σ)λ]2} (33)

2logLθσ=θ[logLσ]=λσ(λ+1)ni=1tiλ[1+(ti/σ)λ] (33)

2logLλσ=λ[logLσ]=nσ+1σ(λ+1)ni=1tiλ[1+(ti/σ)λ][(1λlogσ)+λlogti[1(ti/σ)λ][1+(ti/σ)λ]] (34)

2logLσ2=σ[logLσ]=nλσ2+tiλσ(λ+1)[1+(ti/σ)λ]{(λ+1)σ+λ(tiλ)σ(λ+1)[1+(ti/σ)λ]} (35)

The Asymptotic (1α) 100%  confidence interval of (ˆθ,ˆλandˆσ) are ˆθ±Zα2Var(ˆθ) , ˆλ±Zα2Var(ˆλ) and ˆσ±Zα2Var(ˆσ) respectively, where Zα2  is the upper (α2)th percentile of the standard normal distribution.

To obtain the asymptotic confidence interval for ˆR(t) , we proceed as follows.

The asymptotic variance of the MLEs are given by

V(ˆλ)=[E(2Lλ2)]1=E[nλ2+(θ+1)ni=1{(ti/σ)λ[1(ti/σ)λ][log(ti/σ)]2[1+(ti/σ)λ]2}]1            (36)

V(ˆθ)=[E(2Lθ2)]1=E[nθ2]1=θ2n                                           (37)

V(ˆσ)=[E(2Lσ2)]1=E[nλσ2tiλσ(λ+1)[1+(ti/σ)λ]{(λ+1)σ+λ(tiλ)σ(λ+1)[1+(ti/σ)λ]}]1     (38)

Now R(t)λ=θ(1+(t/σ)λ)1θ(t/σ)λlog[t/σ]                                         (39)

R(t)θ=(1+(t/σ)λ)θlog[1+(t/σ)λ]                                          (40)

R(t)σ=tλθ(1+(t/σ)λ)1θ(t/σ)1+λσ2                                           (41)

The asymptotic variance (A V) of an estimate of ˆR(t) which is a function of parameter estimates (say) ˆλ,ˆθandˆσ  is given by Rao15

AV=V(ˆλ)(R(t)λ)2+V(ˆθ)(R(t)θ)2+V(ˆσ)(R(t)σ)2                               (42)

Thus the asymptotic variance of ˆR(t) can be obtained after substituting equations (36) to (41) in equation (42), which will be analytically solved thereafter.

Samples are generated at varying sizes n = 20, 30, 40, 50 and 100 for different assumed values of shape parameters λ and θ when σ=30 . MLEs, mean square error (MSE) and bias of λ when θ is assumed as 1.5, 2 and 2.5 are presented in Table 2. Similarly MLEs, MSE and bias of θ when λ is assumed as 1.5, 2 and 2.5 are presented in Table 3. Table 4 represents the values of MLEs, mean square error (MSE) and bias of reliability function (R) for the parametric combinations of (λ, θ) = (1.5, 1.5), (1.5, 2), (2, 1.5), (2, 2), (2, 2.5), (2.5, 2) and (2.5, 2.5).

 

n = 20

n = 30

n = 40

n = 50

n = 100

(θ = 2)

         

λ

1.5000

1.5000

1.5000

1.5000

1.5000

ˆλ

1.5968

1.5625

1.5471

1.5372

1.5182

Bias

0.0968

0.0625

0.0471

0.0372

0.0182

MSE

0.0991

0.0576

0.0410

0.0319

0.0148

(θ =1.5)

         

λ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆλ

2.1374

2.0885

2.0668

2.0527

2.0257

Bias

0.1374

0.0885

0.0668

0.0527

0.0257

MSE

0.2027

0.1166

0.0828

0.0641

0.0297

(θ = 2)

         

λ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆλ

2.1290

2.0833

2.0628

2.0496

2.0242

Bias

0.1290

0.0833

0.0628

0.0496

0.0242

MSE

0.1761

0.1025

0.0730

0.0566

0.0263

(θ =2.5)

         

λ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆλ

2.1264

2.0816

2.0614

2.0485

2.0237

Bias

0.1264

0.0816

0.0614

0.0485

0.0237

MSE

0.1648

0.0959

0.0682

0.0530

0.0246

(θ =2.5)

         

λ

2.5000

2.5000

2.5000

2.5000

2.5000

ˆλ

2.6580

2.6019

2.5768

2.5606

2.5296

Bias

0.1580

0.1019

0.0768

0.0606

0.0296

MSE

0.2574

0.1499

0.1066

0.0828

0.0385

Table 2 MLE of λ when σ = 30 for TGLLD

 

n = 20

n = 30

n = 40

n = 50

n =100

(λ = 1.5)

         
θ

1.5000

1.5000

1.5000

1.5000

1.5000

ˆθ

1.5771

1.5487

1.5347

1.5279

1.5133

Bias

0.0771

0.0487

0.0347

0.0279

0.0133

MSE

0.1664

0.0962

0.0679

0.0524

0.0246

(λ = 1.5)

         
θ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆθ

2.1380

2.0863

2.0619

2.0494

2.0235

Bias

0.1380

0.0863

0.0619

0.0494

0.0235

MSE

0.3282

0.1794

0.1236

0.0940

0.0433

(λ = 2)

         
θ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆθ

2.1380

2.0863

2.0619

2.0494

2.0235

Bias

0.1380

0.0863

0.0619

0.0494

0.0235

MSE

0.3282

0.1794

0.1236

0.0940

0.0433

(λ = 2.5)

         
θ

2.0000

2.0000

2.0000

2.0000

2.0000

ˆθ

2.1380

2.0863

2.0619

2.0494

2.0235

Bias

0.1380

0.0863

0.0619

0.0494

0.0235

MSE

0.3282

0.1794

0.1236

0.0940

0.0433

(λ = 2)

         
θ

2.5000

2.5000

2.5000

2.5000

2.5000

ˆθ

2.7155

2.6334

2.5959

2.5761

2.5361

Bias

0.2155

0.1334

0.0959

0.0761

0.0361

MSE

0.6020

0.3118

0.2096

0.1575

0.0708

(λ = 2.5)

         
θ

2.5000

2.5000

2.5000

2.5000

2.5000

ˆθ

2.7155

2.6334

2.5959

2.5761

2.5361

Bias

0.2155

0.1334

0.0959

0.0761

0.0361

MSE

0.6020

0.3118

0.2096

0.1575

0.0708

Table 3 MLE of θ when σ = 30 for TGLLD

 

n = 20

n = 30

n = 40

n = 50

n =100

(λ = 1.5, θ = 1.5)

         
ˆR

0.9977

0.9977

0.9977

0.9977

0.9977

Bias

-0.0009

-0.0006

-0.0004

-0.0004

-0.0002

MSE

0.0005

0.0004

0.0004

0.0003

0.0002

(λ = 1.5, θ = 2)

         
ˆR

0.9978

0.9978

0.9978

0.9978

0.9978

Bias

-0.0008

-0.0005

-0.0004

-0.0003

-0.0002

MSE

0.0008

0.0007

0.0007

0.0004

0.0002

(λ = 2, θ = 1.5)

         
ˆR

0.9996

0.9996

0.9996

0.9996

0.9996

Bias

0.0003

-0.0002

-0.0002

-0.0001

-0.0001

MSE

0.0008

0.0006

0.0005

0.0003

0.0002

(λ = 2, θ = 2)

         
ˆR

0.9996

0.9996

0.9996

0.9996

0.9996

Bias

-0.0003

-0.0002

-0.0001

-0.0001

-0.0001

MSE

0.0009

0.0007

0.0006

0.0005

0.0003

(λ = 2, θ = 2.5)

         
ˆR

0.9996

0.9996

0.9996

0.9996

0.9996

Bias

-0.0003

-0.0002

-0.0001

-0.0001

0.0000

MSE

0.0009

0.0008

0.0007

0.0006

0.0003

(λ = 2.5, θ = 2)

         
ˆR

0.9999

0.9999

0.9999

0.9999

0.9999

Bias

-0.0001

-0.0001

0.0000

0.0000

0.0000

MSE

0.0009

0.0008

0.0006

0.0005

0.0003

(λ = 2.5, θ = 2.5)

         
ˆR

0.9999

0.9999

0.9999

0.9999

0.9999

Bias

-0.0001

-0.0001

0.0001

0.0001

0.0002

MSE

0.0009

0.0008

0.0007

0.0006

0.0004

Table 4 MLE of R when σ = 30 for TGLLD

It can be seen from Tables 2 and 3, the bias and MSE of both λ and θ decrease as the sample size increases. This reflects the asymptotic consistency property of the MLEs. From Table 2, when the parameter λ increases the concerned bias and MSE are also observed as increases and the rate of this increment is high at smaller sample sizes especially till sample size n reaches 50. Similar pattern noticed for the parameter θ from Table 3.

From Table 4, it is observed that when both parameters are increases, the ˆR(t) resulted more accurate values and the absolute bias decreases as the sample size increases. Also mse of ˆR(t) observed as very small values. From Table 5 we noticed that average length of 95% confidence interval is decreases as sample size and parametric values are increases furthermore the coverage probabilities for λ and θ are very close to 95% that we expected. The results shows that estimation and confidence limits of shape parameter λ and θ are performed well using maximum likelihood estimation.

n

λ

θ

Average length

Coverage Probability

λ

θ

λ

θ

20

2.0

1.5

1.5510

1.4140

0.9410

0.9373

30

2.0

1.5

1.2360

1.1310

0.9455

0.9423

40

2.0

1.5

1.0590

0.9701

0.9461

0.9433

50

2.0

1.5

0.9403

0.8634

0.9471

0.9449

100

2.0

1.5

0.6559

0.6043

0.9474

0.9484

20

1.5

1.5

1.1640

1.4140

0.9410

0.9373

30

1.5

1.5

0.9270

1.1310

0.9456

0.9423

40

1.5

1.5

0.7942

0.9701

0.9461

0.9433

50

1.5

1.5

0.7052

0.8634

0.9470

0.9449

100

1.5

1.5

0.4919

0.6043

0.9474

0.9485

20

2.0

2.0

1.4550

1.9070

0.9406

0.9375

30

2.0

2.0

1.1620

1.5080

0.9453

0.9422

40

2.0

2.0

0.9965

1.2870

0.9463

0.9434

50

2.0

2.0

0.8855

1.1420

0.9471

0.9445

100

2.0

2.0

0.6185

0.7951

0.9479

0.9476

20

2.5

2.0

1.8180

1.9070

0.9406

0.9375

30

2.5

2.0

1.4520

1.5080

0.9453

0.9422

40

2.5

2.0

1.2460

1.2870

0.9463

0.9434

50

2.5

2.0

1.1070

1.1420

0.9471

0.9445

100

2.5

2.0

0.7732

0.7951

0.9479

0.9476

20

2.5

2.5

1.7570

2.4850

0.9403

0.9367

30

2.5

2.5

1.4040

1.9410

0.9449

0.9418

40

2.5

2.5

1.2040

1.6480

0.9461

0.9440

50

2.5

2.5

1.0700

1.4580

0.9470

0.9432

100

2.5

2.5

0.7476

1.0100

0.9481

0.9482

20

1.5

2.0

1.0910

1.9070

0.9406

0.9375

30

1.5

2.0

0.8715

1.5080

0.9453

0.9422

40

1.5

2.0

0.7474

1.2870

0.9463

0.9434

50

1.5

2.0

0.6641

1.1420

0.9471

0.9445

100

1.5

2.0

0.4639

0.7951

0.9479

0.9476

20

2.0

2.5

1.4060

2.4850

0.9403

0.9367

30

2.0

2.5

1.1230

1.9410

0.9449

0.9418

40

2.0

2.5

0.9633

1.6480

0.9461

0.9440

50

2.0

2.5

0.8560

1.4580

0.9471

0.9432

100

2.0

2.5

0.5980

1.0100

0.9482

0.9482

Table 5 Average length and coverage probability of λ  and θ  when σ =30

Fitting reliability data

In this section, we considered a real data set to compare the ML estimates of TGLLD given in (3), with the log-logistic (LL), McDonald log-logistic (McLL) studied by Tahir et al.,16 beta log-logistic (BeLL) by Lemonte,17 Kumaraswamy log-logistic (KwLL) discussed by de Santana et al.,18 Marshal-Olkin log-logistic (MoLL) developed by Gui.19 The corresponding densities functions of the above distributions are reproduced below.

LL: f(t)=(λσ)(t/σ)λ1[1+(t/σ)λ]2,t>0,σ>0,λ>1              

McLL: f(t)=cB(ac1,b)(λ/σ)(t/σ)aλ1[1+(t/σ)λ](a+1)[1{1[1+(t/σ)λ]1}c]b1,t>0

BeLL:  f(t)=1B(a,b)(λ/σ)(t/σ)aλ1[1+(t/σ)λ](a+b)  

KwLL: f(t)=ab(λ/σ)(t/σ)aλ1[1+(t/σ)λ](a+1)[1{111+(t/σ)λ}a]b1

MoLL: f(t)=(αλ/σ)(t/σ)λ1[α+(t/σ)λ]2,0t;α,σ>0;λ>1

We will be considering the single fibers of 20mm in gauge length, with sample sizes n = 69 respectively. For easy reference, this data is presented below.

1.312, 1.314, 1.479, 1.552, 1.700, 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, 2.140, 2.179, 2.224, 2.240, 2.253, 2.270, 2.272, 2.274, 2.301, 2.301, 2.359, 2.382, 2.382, 2.426, 2.434, 2.435, 2.478, 2.490, 2.511, 2.514, 2.535, 2.554, 2.566, 2.570, 2.586, 2.629, 2.633, 2.642, 2.648, 2.684, 2.697, 2.726, 2.770, 2.773, 2.800, 2.809, 2.818, 2.821, 2.848, 2.880, 2.954, 3.012, 3.067, 3.084, 3.090, 3.096, 3.128, 3.233, 3.433, 3.585, 3.585.

Table 6 describes the fitted data of MLEs of the unknown parameters of different log-logistic distributions, it is observed that TGLLD throws better results and found more suitable in analyzing the data. Plots of empirical and fitted PDFs drawn for the observed data are shown in Figure 5.

Model

Estimates

-2logL

AIC

BIC

TGLLD (λ, σ, θ)

6.509 (1.0967)

3.119 (0.6536)

3.595 (3.5802)

   

97.84

103.8

110.5

LL (λ, σ)

8.484 (0.8531)

2.431 (0.0598)

     

101.56

105.6

110

McLL (λ, σ, a, b, c)

5.493 (18.802)

2.655 (0.07716)

1.282
(5.6039)

4.353 (30.5138)

4.42 (11.27)

97.82

107.8

119

BeLL (λ, σ, a, b)

5.624 (5.441)

3.298
(1.755)

1.227
(1.617)

5.129 (14.521)

 

97.78

105.8

114.7

KwLL (λ, σ, a, b)

4.002 (10.091)

3.436
(1.811)

1.778
(5.328)

11.064 (70.15)

 

97.74

105.7

114.7

MoLL (λ, σ, α)

8.484 (0.8532)

2.055 (2.3428)

4.162 (40.2466)

 

 

101.56

 

107.6

 

114.3

 

Table 6 The MLEs (SEs in parentheses) and statistics of the distribution parameters

Figure 5 Fitted PDFs of different log-logistic distributions along with TGLLD and Q-Q plot for the observed data.

Conclusion

In this article, a lifetime distribution named as Type-II generalized log-logistic distribution (TGLLD) is considered. We obtained the properties of the distribution viz. mean, percentile, median, quantile function, moments, variance, skewness, kurtosis and distribution function of ith order statistics. Also derived the MLEs of unknown parameters, reliability function and obtained their asymptotic variances. The estimated values are presented through simulation. With the use of real data, the model values are validated and the results compared with other log-logistic distributions by finding the statistics such as -2logL, AIC and BIC. With observed results, it is noticed that this model proven to be more suitable for lifetime models.

Authors’ contributions

GSR contributed to study design, data analysis and drafting of the manuscript. SV and KR contributed to this work by data acquisition and revising of the manuscript. All co-authors have given the final approval of the version to be published.

Funding

This study did not receive any funding in any form.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests. The authors alone are responsible for the content and the writing of the paper.

Acknowledgment

None.

References

  1. Rosaiah K, Kantam RRL, Prasad SVSVSV, et al. Reliability Estimation in Type-II Generalized Log-Logistic distribution, International Journal of Agricultural and Statistical Sciences. 2008;4(2):283‒292.
  2. Balakrishnan N. Order Statistics from the half logistic distribution, Journal of Statistical Computation and Simulation. 1985;20(4):287‒309.
  3. Balakrishnan N. Approximate maximum likelihood estimation for a generalized logistic distribution. Journal of Statistical Planning and inference. 1990;26(2):221‒236.
  4. Mudholkar GS, Srivastava DK. Exponentiated Weibull family for analyzing bathtub failure-rate data, IEEE Transactions Rehabilitation Engineering. 1993;42:299‒302.
  5. Mudholkar GS, Srivastava DK, Freimer M. The exponentiated Weibull family: A Reanalysis of bus-motor-failure data, 1995;37(4):436‒445.
  6. Gupta RD, Kundu D. Generalized exponential distribution, Australian and New Zealand Journal of Statistics. 1999;41(2):173‒188.
  7. Nadarajah S. The exponentiated Gumbel distribution with climate application, 2005;17(1):13‒23.
  8. Nadarajah S, Gupta AK. The exponentiated Gamma distribution with application to drought data, Calcutta Statistical Association. 2007;59:29‒54.
  9. Abouammoh AM, Alshingiti AM. Reliability estimation of generalized inverted exponential distribution. Journal of Statistical Computation and Simulation. 2009;79(11):1301‒1315.
  10. Rosaiah K, Rao GS, Kalyani K, et al. Odds Exponential Log Logistic Distribution: Properties and Estimation. Journal of Mathematics and Statistics. 2017;13(1):14‒23.
  11. Ashkar F, Mahdi S. Fitting the log-logistic distribution by generalized moments, Journal of Hydrology. 2006;328(3-4):694‒703.
  12. Rosaiah K, Kantam RRL, Santosh KumarCh. Reliability test plans for exponentiated log-logistic distribution, Economic Quality Control. 2006;21(2):279‒289.
  13. Rao GS, Kantam RRL, Rosaiah K, et al. Reliability test plans for type-II generalized log-logistic distribution, Journal of Reliability and Statistical Studies. 2012;5(1):55‒64.
  14. Rao GS, Kantam RRL, Rosaiah K et al. An economic reliability test plan for generalized log-logistic distribution, International Journal of Engineering and Applied Sciences. 2012;3:61‒68.
  15. Rao CR. Linear Statistical Inference and its Applications: Second Editon. John Wiley & Sons, Inc.
  16. Tahir MH, Muhammad M, Muhammad Z, et al. McDonald log-logistic distribution with an application to breast cancer data. Journal of Statistical Theory and Applications. 13(1):65-82.
  17. Lemonte AJ. The beta log-logistic distribution. Brazilian Journal of Probability and Statistics. 2014;28:313-332.
  18. de Santana TVF, Ortega EMM, Cordeiro GM, et al. The Kumaraswamy-logLogistic distribution. Journal of Statistical Theory and Applications. 11:265–291.
  19. Gui W. Marshall-Olkin extended log-logistic distribution and its application in minification processes. Applied Mathematical Sciences. 2013;7:3947-3961.
Creative Commons Attribution License

©2022 Prasad, 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.