Research article Special Issues

Bivariate q-extended Weibull morgenstern family and correlation coefficient formulas for some of its sub-models

  • A bivariate extension of a flexible univariate family is proposed. The new family is called bivariate q-extended Weibull Morgenstern family of distributions which can be constructed based on the Farlie-Gumbel-Morgenstern (FGM) copula technique. After introducing the new family, four sub-models are discussed in detail from the theoretical and numerical coefficient of correlation point of view with pointing to the effect of the q parameter.

    Citation: Rasha Abd-Elwahaab Attwa, Taha Radwan, Esraa Osama Abo Zaid. Bivariate q-extended Weibull morgenstern family and correlation coefficient formulas for some of its sub-models[J]. AIMS Mathematics, 2023, 8(11): 25325-25342. doi: 10.3934/math.20231292

    Related Papers:

    [1] M. Nagy, H. M. Barakat, M. A. Alawady, I. A. Husseiny, A. F. Alrasheedi, T. S. Taher, A. H. Mansi, M. O. Mohamed . Inference and other aspects for $ q- $Weibull distribution via generalized order statistics with applications to medical datasets. AIMS Mathematics, 2024, 9(4): 8311-8338. doi: 10.3934/math.2024404
    [2] H. M. Barakat, M. A. Alawady, I. A. Husseiny, M. Nagy, A. H. Mansi, M. O. Mohamed . Bivariate Epanechnikov-exponential distribution: statistical properties, reliability measures, and applications to computer science data. AIMS Mathematics, 2024, 9(11): 32299-32327. doi: 10.3934/math.20241550
    [3] Aisha Fayomi, Ehab M. Almetwally, Maha E. Qura . A novel bivariate Lomax-G family of distributions: Properties, inference, and applications to environmental, medical, and computer science data. AIMS Mathematics, 2023, 8(8): 17539-17584. doi: 10.3934/math.2023896
    [4] Jumanah Ahmed Darwish, Saman Hanif Shahbaz, Lutfiah Ismail Al-Turk, Muhammad Qaiser Shahbaz . Some bivariate and multivariate families of distributions: Theory, inference and application. AIMS Mathematics, 2022, 7(8): 15584-15611. doi: 10.3934/math.2022854
    [5] Ayed. R. A. Alanzi, M. Qaisar Rafique, M. H. Tahir, Farrukh Jamal, M. Adnan Hussain, Waqas Sami . A novel Muth generalized family of distributions: Properties and applications to quality control. AIMS Mathematics, 2023, 8(3): 6559-6580. doi: 10.3934/math.2023331
    [6] Sajid Mehboob Zaidi, Mashail M. AL Sobhi, M. El-Morshedy, Ahmed Z. Afify . A new generalized family of distributions: Properties and applications. AIMS Mathematics, 2021, 6(1): 456-476. doi: 10.3934/math.2021028
    [7] Qasim Ramzan, Muhammad Amin, Ahmed Elhassanein, Muhammad Ikram . The extended generalized inverted Kumaraswamy Weibull distribution: Properties and applications. AIMS Mathematics, 2021, 6(9): 9955-9980. doi: 10.3934/math.2021579
    [8] Areej M. AL-Zaydi . On concomitants of generalized order statistics arising from bivariate generalized Weibull distribution and its application in estimation. AIMS Mathematics, 2024, 9(8): 22002-22021. doi: 10.3934/math.20241069
    [9] Walid Emam . Benefiting from statistical modeling in the analysis of current health expenditure to gross domestic product. AIMS Mathematics, 2023, 8(5): 12398-12421. doi: 10.3934/math.2023623
    [10] Mohammed Albassam, Muhammad Ahsan-ul-Haq, Muhammad Aslam . Weibull distribution under indeterminacy with applications. AIMS Mathematics, 2023, 8(5): 10745-10757. doi: 10.3934/math.2023545
  • A bivariate extension of a flexible univariate family is proposed. The new family is called bivariate q-extended Weibull Morgenstern family of distributions which can be constructed based on the Farlie-Gumbel-Morgenstern (FGM) copula technique. After introducing the new family, four sub-models are discussed in detail from the theoretical and numerical coefficient of correlation point of view with pointing to the effect of the q parameter.



    Let (X,Y) be a pair of absolutely continuous random variables (r.v.'s), formally defined by the cumulative distribution function (DF)

    FX,Y(x,y)=FX(x)FY(y){1+λA(FX(x))B(FY(y))},

    where FX(x) and FY(y) are the marginals DF's of X and Y, respectively. Moreover, the two kernels A(x)0 and B(x)0, as x1, satisfy certain regularity conditions ensuring that FX,Y(x,y) is a DF with absolutely continuous marginals FX(x) and FY(y). Usually, any bivariate distribution FX,Y(x,y) that has the above representation is referred to as FGM distribution. This model was originally introduced by Morgenstern (1956)[11] for A(x)=1x and B(y)=1y and investigated by Gumbel (1960)[7] for exponential with its multivariate case. The admissible range of association parameter λ for the distribution, with A(x)=1x and B(y)=1y, is 1λ1 and the Pearson correlation coefficient between X and Y can never exceed 1/3. The classical FGM distribution is a flexible family useful in applications provided that the correlation between the variables is not too large. It can be utilized for arbitrary continuous marginals FX(x) and FY(y).

    FX,Y(x,y)=FX(x)FY(y)[1+λ(1FX(x))(1FY(y))],1λ1, (1.1)
    fX,Y(x,y)=fX(x)fY(y)[1+λ(2FX(x)1)(2FY(y)1)],1λ1. (1.2)

    The maximum correlation coefficient is 1/4 when the marginal distributions are exponential (Gumbel, 1960[7]). The maximum correlation coefficients equal 0.2921 and 0.314 for generalized exponential (Tahmasebi and Jafari, 2015[17]) and Rayleigh distributions (Tahmasebi et al., 2018[18]), restrictively.

    In mathematical physics and probability, the q-distribution is more general than classical distribution. It was introduced by Diaz and Parguan (2009)[4] and Diaz et al. (2010)[3] in the continuous case and by Charalambides (2010)[1] in the discrete case. The construction of a q-distribution is the construction of a q-analogue of ordinary distribution. Mathai (2005)[10] introduced the q-analogue of the gamma distribution with respect to Lebesgue measure. Recently, several q-type super statistical distributions, such as the q-exponential, q-Weibull and q-logistic, were developed in the context of statistical mechanics, information theory and reliability modelling as discussed for instance by Sol and Kac (2005)[15] and Srivastave and Choi (2012)[16]. Provost et al. (2018)[14] introduced the q-generalized extreme value under linear normalization (q-GEVL) and q-Gumbel distributions.

    Gurvich et al. (1997)[8] introduced a class of extended Weibull family (EWF) of distributions with survival function (SF),

    ˉF(x;θ,η)=eθΨ(x;η);x>0,θ>0, (1.3)

    where η is a vector of parameters and Ψ(x;η) is a non-negative, continuous, monotone increasing and differentiable function of x, which depends on the parameter vector η. Also, Ψ(x;η)0+ as x0+ and Ψ(x;η) as x. The probability density function (PDF) corresponding to (1.3) can be listed as

    f(x;θ,η)=θΨ(1)(x;η)eθΨ(x;η),x>0;θ>0, (1.4)

    where Ψ(1)(x,η) is the derivative of Ψ(x;η) with respect to x.

    Let X be a continuous random variable from the new family which is introduced by Nigm and Sadk (2021a). This family is called q-extended Weibull family and denoted by q-EWF. The SF is:

    ˉF(x;q,θ,η)=[1+qθΨ(x;η)]1q;x>0,θ>0,q0, (1.5)

    where η could be a vector of parameters and Ψ(x;η) is non-negative, monotonically increasing function of x, not depending on θ and differentiable with respect to x. The PDF corresponding to (1.5) is:

    f(x;q,θ,η)=θΨ(1)(x;η)[1+qθΨ(x;η)](1+q)q;x>0,θ>0,q0. (1.6)

    Note that if q0 in (1.5) and (1.6) we have (1.3) and (1.4).

    Some special cases of the q-EWF with the corresponding Ψ(x;η) are listed in Table 1.

    Table 1.  Special distributions of the q-EWF.
    Distribution Ψ(x;η) Support θ η
    q- Exponential x x0 θ
    q- Pareto log(xb) xb θ b
    q- Weibull xb x0 θ b
    q- Frechet log(1exb) x0 1 b
    q- Lomax log(1+xb) x0 θ b
    q- Linear failure rate ax+bx22 x0 1 (a,b)
    q- Burr XII log(1+(xc)b) x0 θ (b,c)

     | Show Table
    DownLoad: CSV

    The inference for q-EWF were considered under Type-II progressive censoring with random removal by Nigm and Sadk (2021b)[12]. Also, a class of models called Marshall-Olkin q-extended Weibull family was proposed by Nigm and Sadk (2021c)[13]. There are a lot of papers on the same topic (Cousineau (2009)[2], Gauchman (2004)[5], Gradshteyn and Ryzhik (2007)[6] and Krugman (1996)[9]).

    In this paper, we propose a bivariate extension of the q-EWF using the FGM approach. The main topic is finding the maximum correlation coefficient for some special cases of the FGM q-EWF. For some distributions, the correlation coefficient can be obtained mathematically when the rth moment of the marginal distribution exists. Otherwise, the correlation coefficient can be obtained numerically. The paper is organized as follows: In Section 2, some properties of the FGM q-EWF distribution are investigated. In Section 3, some special cases of this family are studied and the maximum correlation is obtained. Finally, some conclusion remarks are reported in Section 4.

    The joint CDF and PDF of (X,Y) are defined by (1.1) and (1.2), respectively, where Xq-EWF(θ1,η1,q1) and Yq-EWF(θ2,η2,q2). Thus, it is easy to show that the (n,m)th joint moment of the FGM q-EWF (λ,θ1,θ2,q1,q2) family is given by

    E(XnYm)=E(Xn)E(Ym)+λ(E(Xn)E(Un1))(E(Ym)E(Um2));n,m=1,2,..., (2.1)

    where U1q-EWF(2θ1;η1,q1/2) and U2q-EWF(2θ2;η2,q2/2).

    Also, the covariance of (X,Y) is

    Cov(X,Y)=λ(E(X)E(U1))(E(Y)E(U2)). (2.2)

    The correlation coefficient of (X,Y) is

    ρ(X,Y)=λ[E(X)E(U1)]Var(X)[E(Y)E(U2)]Var(Y). (2.3)

    In this section, we obtain the maximal correlation coefficient for special cases of the FGM q-EW(θ,η,q) family distribution through the following sub-sections:

    The q-U(θ,η,q) model can be expressed by putting Ψ(x,η1)=log(1x);0<x<1,θ=1 in (1.5) and (1.6). Figure 1 shows the PDF, SF and Hazard function (HF) of q-U(θ,η,q) for various values of the parameter q. According to Figure 1, it is noted that q-U (θ,η,qi) cannot be used to model left and right skewed data.

    Figure 1.  (a) The PDF, (b) SF and (c) HF of q-U(θ,η,q) distribution for different values of parameter q.

    The following theorem lists the new created model based on the coefficient of correlation.

    Theorem 3.1. Let (X,Y) bivariate FGM q-U (θi,ηi,qi), for Ψ(x,η1)=log(1x),Ψ(y,η2)=log(1y),0<x,y<1,θi=1, and Ui qU(2θi,ηi,qi/2),i=1,2. Then

    E(Xn)=nk=0(1)kCnkφk(q1,t)andE(Ym)=mj=0(1)jCmjφj(q2,t), (3.1)
    E(Un1)=2nk=0(1)kCnkξk(q1,t)andE(Um2)=2mj=0(1)jCmjξj(q2,t), (3.2)

    where

    φk(q1,t)=0ekt[1+q1t]1q11dt,φj(q2,t)=0ejt[1+q2t]1q21dt, (3.3)
    ξk(q1,t)=0ekt[1+q1t]2q11dtandξj(q2,t)=0ejt[1+q2t]2q21dt (3.4)

    and

    ρ(X,Y)=λ(2ξ1(q1,t)φ1(q1,t))(2ξ1(q2,t)φ1(q2,t))(φ2(q1,t)φ21(q1,t))(φ2(q2,t)φ21(q2,t)). (3.5)

    Proof. With Ψ(x,η1)=log(1x),θ1=1 in (1.5) and (1.6), we have

    ˉF=[1+(q1log(1x))]1q1, 1q1>0; 0x1.

    Then

    E(Xn)=10xn1x[1+(q1log(1x))]1q11dx.

    Put t=log(1x), we get

    E(Xn)=0(1et)n[1+q1t]1q11dt
    =0[nk=0(1)kCnkekt][1+q1t]1q11dt
    =nk=0(1)kCnk0ekt(1+q1t)1q11dt
    =nk=0(1)kCnkφk(q1,t), (3.6)

    where φk(q1,t) defined in (3.3). Similarly, we have

    E(Ym)=mj=0(1)jCmjφj(q2,t). (3.7)

    where φj(q2,t) defined in (3.3). Also, Ui qU(2θi,ηi,qi2),i=1,2 then:

    E(Un1)=2nk=0(1)kCnkξk(q1,t), (3.8)

    where ξk(q1,t) defined in (3.4) and

    E(Um2)=2mj=0(1)jCmjξj(q2,t), (3.9)

    where ξj(q2,t) defined in (3.4).

    From (3.6) and (3.7) for n=m=1,2, we get

    E(X)=1φ1(q1,t),E(Y)=1φ1(q2,t), (3.10)
    E(X2)=12φ1(q1,t)+φ2(q1,t),E(Y2)=12φ1(q2,t)+φ2(q2,t) (3.11)
    Var(X)=φ2(q1,t)φ21(q1,t),Var(Y)=φ2(q2,t)φ21(q2,t). (3.12)

    From (3.8) and (3.9) for n=m=1 we get

    E(U1)=12ξ1(q1,t),E(U2)=12ξ1(q2,t). (3.13)

    From (3.10), (3.12) and (3.13) in (2.3) we get

    ρ(X,Y)=λ(2ξ1(q1,t)φ1(q1,t))(2ξ1(q2,t)φ1(q2,t))(φ2(q1,t)φ21(q1,t))(φ2(q2,t)φ21(q2,t)). (3.14)

    Remark 3.1. If qi0,i=1,2 we see that φ1(qi,t)=12, φ2(qi,t)=13, ξ1(qi,t)=13 and ρ(X,Y)=λ3.

    The following table (see Table 2) gives the values of the correlation coefficient in case FGM q-Uniform distribution.

    Table 2.  The values of the correlation coefficient in case FGM q-Uniform distribution.
    q1 q2 ρ
    16.9 19.9 0.141
    16.8 19.8 0.1415
    16.6 19.6 0.1424
    2.2 5.2 0.273
    0.5 0.01 0.1212
    -0.01 -0.01 0.333
    -0.02 -0.02 0.333
    0.1 0.1 0.333
    0.5 0.5 0.3331
    -0.1 -0.1 0.333
    -0.5 -0.01 0.3323
    -0.05 -0.01 0.333
    -0.01 -0.001 0.333

     | Show Table
    DownLoad: CSV

    By putting Ψ(x,η)=x,x>0 in (1.5), we will have qExp(θ,η,q). Figure 2 illustrates the PDF, the SF and HF for q-exponential distribution for q=0.3 and different values of θ as follow:

    Figure 2.  (a) The PDF, (b) SF and (c) HF of q-exponential distribution for q=0.3 and different values of θ.

    We obtain the maximum correlation coefficient from special cases of FGM q-EWF (θ,η,q) distribution [qExp(θ,η,q)] through the following theorem.

    Theorem 3.2. Let (X,Y) bivariant FGM q-Exp (θi,ηi,qi), for Ψ(x,η1)=x,Ψ(y,η2)=y,0<x,y<,θi>0 ,Uiq-Exp(2θi,ηi,qi2),i=1,2. Then

     E(Xn)=n!θn1nj=1(1jq1)andE(Ym)=m!θm2mk=1(1kq2),jq1,kq21 (3.15)
     E(Un1)=2n!θn1n+1j=1(2jq1)andE(Um2)=2m!θm2m+1k=1(2kq2),jq1,kq22 (3.16)

    and

    ρ(X,Y)=λ(1q1)(1q2)(12q1)(12q2)(2q1)(2q2),q12,q22. (3.17)

    Proof. With Ψ(x,η1)=x in (1.5) and (1.6) we have,

    ¯F =[1+qθ1x]1q1,q10,0Xi.

    Then

    E(Xn)=0θ1xn[1+q1θ1x)]1q11dx.

    Put t=q1θ1x

    we get

    E(Xn)=0q(n+1)1θ1ntn[1+t]1q11dt.

    Let u=t/(t+1)

    then,

    E(Xn)=10q(n+1)1θn1un[1u]n1+1q1du=q(n+1)1θn1β(n+1,1q1n)
    =n!θn1nj=1(1jq1),jq11. (3.18)

    Similarly, we have

    E(Ym)=m!θn2mk=1(1kq2),kq21. (3.19)

    Also, Uiq-Exp(2θi,ηi,qi/2),i=1,2

    then:

    E(Un1)=2n!θn1n+1j=1(2jq1),jq12 (3.20)

    and

    E(Um2)=2m!θm2m+1k=1(2kq2),kq22. (3.21)

    From (3.18) and (3.19) for n=m=1,2 we get

    E(X)=1θ1(1q1),E(Y)=1θ2(1q2), (3.22)
    E(X2)=2θ21(1q1)(12q1),E(Y2)=2θ22(1q2)(12q2)
    Var(X)=1θ21(1q1)2(12q1),Var(Y)=1θ22(1q2)2(12q2). (3.23)

    From (3.20) and (3.21) for n=m=1 we get

    E(U1)=1θ1(1q1)(2q1),E(U2)=1θ2(1q2)(2q2),qi1,qi2,i=1,2. (3.24)

    From (3.22)–(3.24) in (2.3) we get

    ρ(X,Y)=λ(1q1)(1q2)(12q1)(12q2)(2q1)(2q2)qi1,qi2,i=1,2. (3.25)

    Remark 3.2. If qi0,i=1,2, we see that E(X)=1/θ1,E(Y)=1/θ2,Var(X)=1/θ21,Var(Y)=1/θ22,E(U1)=1/2θ1,E(U2)=1/2θ2andρ(X,Y)=λ/4

    The following tables (see Table 3) gives the values of the correlation coefficient in case FGM q-exponential distribution.

    Table 3.  (a) and (b) show the values of the correlation coefficient in case FGM q-exponential distribution.
    q1 q2 ρ q1 q2 ρ
    -20 -20 0.0847 -2 -6 0.2519456
    -19 -19 0.0884 -2 -5 0.264864
    -18 -18 0.0925 -2 -4 0.2795085
    -17 -17 0.09695 -2 -3 0.295804
    -16 -16 0.1019 -2 -2 0.3125
    -15 -15 0.1073 -2 -1 0.3227
    -14 -14 0.1133 -2 -14 0.1881499
    -13 -13 0.12 -2 0 0.2795085
    -12 -12 0.1276 -2 10 0
    -11 -11 0.1361 -1 -20 0.168038
    -10 -10 0.1458 -1 -19 0.1716929
    -9 -9 0.157 0 -4 0.25
    -8 -8 0.17 0 -3 0.264575
    -7 -7 0.1852 0 -2 0.2795085
    -6 -6 0.203125 0 -1 0.288675
    -5 -5 0.2244898 1 6 -0.829156
    -4 -4 0.25 1 7 -0.7211
    -3 -3 0.28 1 8 -0.645497
    -2 -2 0.3125 1 9 -0.589015
    -1 -1 0.3333 1 10 -0.54486
    0 0 0.25 3 0 0
    6 6 0.6875 8 1 -0.645497
    7 7 0.52 8 4 0.8539126
    8 8 0.416667 8 5 0.645497
    9 9 0.3469388 8 6 0.535218
    10 10 0.296875 8 7 0.4654747
    -18 4 0 4 7 0.953939
    -15 -4 0.1637578 4 8 0.8539126
    -5 -2 0.264842 4 9 0.7791937
    -5 -1 0.2735506 4 10 0.720785
    -4 -4 0.25 5 -20 0
    -3 -16 0.1688743 9 1 -0.589015
    -3 -4 0.2020726 9 4 0.7791937
    -3 -7 0.22771 9 5 0.589015
    -3 -4 0.264575 9 6 0.4883855
    -3 -3 0.28 9 7 0.4247448
    -3 -1 0.3055 9 8 0.3802076
    -2 -20 0.1627025 9 10 0.320932
    -2 -13 0.1936492
    -2 -12 0.1996489
    -2 -11 0.206227
    -2 -10 0.213478
    -2 -7 0.2405626
    (a) (b)

     | Show Table
    DownLoad: CSV

    By putting Ψ(x,η)=xb,x>0 in (1.5), we have q-Weibull(θ,η,q). Figure 3 illustrates the probability density function, survival function, hazard function of q-Weibull distribution for q=0.3 and some different values of θ and b as follows:

    Figure 3.  (a) The PDF, (b) SF and (c) HF of q-Weibull distribution forq=0.3 and some different values of θ and b.

    From Figures 2 and 3, we see that the q-exponential and q-Weibull distributions can be used for modeling various data in different fields. The previous line of the PDF of the two distributions can be unimodal and bimodal. The hazard function for these distributions have various shapes, including increasing, decreasing, unimodal or bathtub.

    The maximum correlation coefficient from special cases of FGM q-EWF (λ,θ,q) family distributions [qW(θ,η,q)] will be obtained through the following theorem.

    Theorem 3.3. Let (X,Y) bivariant FGM q-W (θi,ηi,qi), for, Ψ(x,η1)=xb1,Psi(y,η2)=yb2,0<x,y<,bi>0,θi>0,Ui q-W (2θi,ηi,qi/2),i=1,2. Then

    E(Xn)=β(nb1+1,1q1nb1)θnb11qnb1+11andE(Ym)=β(mb2+1,1q2mb2)θmb22qmb2+12,qi1,i=1,2 (3.26)
     E(Un1)=2β(nb1+1,2q1nb1)θnb11qnb1+11andE(Um2)=2β(mb2+1,2q2mb2)θmb22qmb2+12,qi1,i=1,2 (3.27)

    and

    ρ(X,Y)=λ[β(1b1+1,1q11b1)2β(1b1+1,2q11b1)]q1β(2b1+1,1q12b1)(β(1b1+1,1q11b1))2
    [β(1b2+1,1q21b2)2β(1b2+1,2q21b2)]q2β(2b2+1,1q22b2)(β(1b2+1,1q21b2))2. (3.28)

    Proof. With Ψ(x,η1)=xb1in(1.5) and (1.6) we have

    ˉF=[1+q1θ1xb1]1q1,q10,0x.

    Then

    E(Xn)=0θ1b1xn+b11[1+q1θ1xb1]1q11dx.

    Put t=q1θ1xb1

    E(Xn)=0q(nb1+1)1θ1nb1tnb1[1+t]1q11dt

    let u=t/t+1

    then

    E(Xn)=10q(nb1+1)θ1nb1unb1[1u]nb11+1q1du
    =β(nb1+1,1q1nb1)θnb11qnb1+11,q10. (3.29)

    Similarly, we have

    E(Yn)=β(mb2+1,1q2mb2)θmb22qmb2+12,q20. (3.30)

    Also, UiqW(2θi,ηi,qi/2),i=1,2

    then:

    E(Un1)=2β(nb1+1,2q1nb1)θnb11qnb1+11, (3.31)

    and

    E(Um2)=2β(nb2+1,2q2mb2)θmb22qmb2+12. (3.32)

    From (3.29) and (3.30) for n=m=1,2. we get

    E(X)=β(1b1+1,1q11b1)θ1b11q1b1+11,E(Y)=β(1b2+1,1q21b2)θ1b22q1b2+12, (3.33)
    E(X2)=β(2b1+1,1q12b1)θ2b11q2b1+11,E(Y2)=β(2b2+1,1q22b2)θ2b22q2b2+12Var(X)=θ2b11q2b121[q1β(2b1+1,1q12b1)[β(1b1+1,1q11b1)]2],
    Var(Y)=θ2b22q2b222[q2β(2b2+1,1q22b2)[β(1b2+1,1q21b2)]2]. (3.34)

    From (3.31) and (3.32) for n=m=1 we get

    E(U1)=2β(1b1+1,2q11b1)θ1b11q1b1+11,E(U2)=2β(1b2+1,2q21b2)θ1b22q1b2+12. (3.35)

    From (3.33)–(3.35) in (2.3) we get

    ρ(X,Y)=λ[β(1b1+1,1q11b1)2β(1b1+1,2q11b1)]q1β(2b1+1,1q12b1)(β(1b1+1,1q11b1))2
    [β(1b2+1,1q21b2)2β(1b2+1,2q21b2)]q2β(2b2+1,1q22b2)(β(1b2+1,1q21b2))2. (3.36)

    Remark 3.3. From Theorem 3.2:

    (1) If qi0,i=1,2 we see that,

    E(X)=θ11b1Γ(1+1b1),E(Y)=θ21b2Γ(1+1b2),Var(X)=θ12b1[Γ(1+2b1)(Γ(1+1b1))2],Var(Y)=θ22b2[Γ(1+2b2)(Γ(1+1b2))2],E(U1)=(2θ1)1b1Γ(1+1b1),E(U2)=(2θ2)1b2Γ(1+1b2)

    then,

    ρ(X,Y)=λ(121b1)(121b2)[Γ(1+1b1)Γ(1+1b2)]Γ(1+2b1)(Γ(1+1b1))2Γ(1+2b2)(Γ(1+1b2))2.

    (2) If b1=b2=1 then q-Weibull distribution becomes q-exponential distribution (as in Theorem 3.2).

    The following table (see Table 4) gives the values of the correlation coefficient in case FGM q-Weibull distribution.

    Table 4.  The values of the correlation coefficient in case FGM q-Weibull distribution for b1=b2=2 in (a) and for b1=b2=3 in (b).
    q1 q2 ρ q1 q2 ρ
    0.01 0.0001 0.31356 0.01 0.002 0.3219998
    0.001 0.0001 0.31389 0.01 0.0021 0.321998
    0.2 1 0 0.01 0.0022 0.3219966
    0.3 0.1 0.2890665 0.01 0.0023 0.321995
    0.3 0.7 0.2064799 0.01 0.0024 0.3219935
    0.3 0.8 0.173983 0.01 0.0082 0.3219
    0.3 0.9 0.126611 0.01 0.0083 0.3218983
    0.3 1 0 0.01 0.0084 0.3218967
    0.4 0.9 0.1215606 0.01 0.0085 0.321895
    0.4 1 0 0.01 0.0086 0.3218934
    0.5 0.1 0.262496 0.01 0.0087 0.3218918
    0.5 0.2 0.2586 0.01 0.0088 0.32189
    0.5 0.3 0.249315 0.4 0.1 0.3053376
    0.9 0.8 0.080194 0.4 0.2 0.3024359
    0.9 0.9 0.058387 0.4 0.3 0.2983908
    0.9 1 0 0.4 0.4 0.2930296
    1 0.1 0 0.4 0.5 0.2861815
    0.9 0.2 0.13044 0.01 0.0089 0.3218885
    0.9 0.3 0.12661 0.01 0.009 0.3218868
    0.9 0.4 0.1215606 0.01 0.0091 0.3218852
    0.9 0.5 0.114973 0.01 0.0092 0.3218835
    0.9 0.6 0.10641 0.01 0.0093 0.3218819
    0.9 0.7 0.09521949 0.01 0.0094 0.3218803
    1 0.2 0 0.01 0.0095 0.3218786
    1 0.3 0 0.01 0.0096 0.321877
    0.01 0.0097 0.3218753
    0.01 0.0098 0.3218737
    0.01 0.0099 0.321872
    0.01 0.01 0.3218704
    0.9 0.1 0.2506464
    0.9 0.2 0.2480838
    0.9 0.3 0.2447657
    0.4 0.8 0.254991
    0.5 0.1 0.2982018
    0.5 0.2 0.2953679
    0.5 0.3 0.291417
    (a) (b)

     | Show Table
    DownLoad: CSV

    By putting Ψ(x,η)=log(x/b) in (1.5) ,η=b,xb we will have qpareto(θ,b,q). Figure 4 illustrates the PDF, the SF and HF for q-pareto distribution for q=1 and different values of θ and b as follow:

    Figure 4.  (a) The PDF, (b) SF and (c) HF of q-pareto distribution for q=1 and different values of θ and b.

    The maximum correlation coefficient from special case of FGM q-pareto (θ,η,q) distribution [qpareto(θ,b,q)] will be obtained through the following theorem.

    Theorem 3.4. Let (X,Y) bivariant FGM q-pareto (θi,bi,qi), for Ψ(x,η1)=log(x/b1),Ψ(y,η2)=log(y/b2),xb1,yb2,θi>0 ,Ui q-pareto (2θi,bi,qi/2),i=1,2. Then

    E(Xn)=bn1k=0nkk!φk(t;q1,θ1),E(Ym)=bm2j=0mjj!φj(t;q2,θ2) (3.37)
    E(Un1)=2bn1k=0nkk!φk(t;q1,θ1),andE(Um2)=2bm2j=0mjj!φj(t;q2,θ2) (3.38)

    where,

    φk(t;q1,θ1)=θ10tk(1+q1θ1t)1q11dt,φj(t;q2,θ2)=θ20tj(1+q2θ2t)1q21dt,

    φk(t;q1,θ1)=θ10tk(1+q1θ1t)2q11dt,φj(t;q2,θ2)=θ20tj(1+q2θ2t)2q21dt,

    and

    ρ(X,Y)=λ[(1+φ1(t;θ1,q1))2(1+φ1(t;θ1,q1))][(1+φ1(t;θ2,q2))2(1+φ1(t;θ2,q2))](2φ2(t;θ1,q1)φ21(t;θ1,q1))(2φ2(t;θ2,q2)φ21(t;θ2,q2)). (3.39)

    Proof. With Ψ(x,η1)=log(xb1) in (1.5) and (1.6) we have,

    ¯F =[1+q1θ1log(xb1)]1q1,,xb1,q10.

    Then

    E(Xn)=b1θ1xn1[1+q1θ1logxb1]1q11dx.

    Put t=logx/b1 we get

    E(Xn)=θ1bn10ent[1+q1θ1t]1q11dt.

    Then,

    E(Xn)=bn1θ10k=0(nt)kk![1+q1θ1t]1q11dt
    =bn1k=0nkk!φk(t;q1,θ1) (3.40)

    where, φk(t;q1,θ1)=θ10tk(1+q1θ1t)1q11dt.

    Similarly, we have

    E(Ym)=bm2j=0mjj!φj(t;q2,θ2) (3.41)

    where, φj(t;q2,θ2)=θ20tj(1+q2θ2t)1q21dt.

    Also, Uiqpareto(2θi,ηi,qi2), then,

    E(Un1)=2bn1k=0nkk!φk(t;q1,θ1) (3.42)

    where, φk(t;q1,θ1)=θ10tk(1+q1θ1t)2q11dt

    and

    E(Um2)=2bm2j=0mjj!φj(t;q2,θ2) (3.43)

    where, φj(t;q2,θ2)=θ20tj(1+q2θ2t)2q21dt.

    From (3.18) and (3.41) for n=m=1,2 we get

    E(X)=b1[1+φ1(t;q1,θ1)],E(Y)=b2[1+φ1(t;q2,θ2)], (3.44)
    E(X2)=b21[1+2φ1(t;q1,θ1)+2φ2(t;q1,θ1)],
    E(Y2)=b22[1+2φ1(t;q2,θ2)+2φ2(t;q2,θ2)]

    then,

    Var(X)=b21[2φ2(t;q1,θ1)φ21(t;q1,θ1)],
    Var(Y)=b22[2φ2(t;q2,θ2)φ21(t;q2,θ2)]. (3.45)

    From (3.19) and (3.43) for n=m=1 we get

    E(U1)=2b1[1+φ1(t;q1,θ1)],E(U2)=2b2[1+φ1(t;q2,θ2)]. (3.46)

    From (3.44)–(3.46) in (2.3) we get

    ρ(X,Y)=λ[(1+φ1(t;θ1,q1))2(1+φ1(t;θ1,q1))][(1+φ1(t;θ2,q2))2(1+φ1(t;θ2,q2))](2φ2(t;θ1,q1)φ21(t;θ1,q1))(2φ2(t;θ2,q2)φ21(t;θ2,q2)). (3.47)

    Remark 3.4. If qi0,i=1,2, we see that

    ρ(X,Y)=λ4(12θ1)(12θ2)(4θ11)(4θ21),θi>1/4,i=1,2

    for θi then |ρ(X,Y)|<14.

    Table 5 gives the values of the correlation coefficient in case FGM q-pareto distribution.

    Table 5.  The values of the correlation coefficient in case FGM q-pareto distribution.
    q1 q2 θ1 θ2 ρ
    0.1 0.1 0.4 0.4 0.0079
    0.01 0.01 0.1 0.1 0.0535
    0.01 0.03 2 2 0.6862
    0.1 0.2 1.7 1.7 0.2073
    0.2 0.2 0.1 0.1 0.0522
    0.1 0.3 2 2 0.2256
    0.16 0.16 3 2 0.57
    0.16 0.01 3 2 0.8436
    0.6 0.05 1 2 0
    0.1 0.1 2.5 2.5 0.8489
    0.1 0.1 2.4 2.4 0.7625
    0.1 0.1 0.1 0.1 0.054
    0.05 0.05 1 1 0.0593
    0.05 0.03 1 1 0
    0.1 0.1 2.6 2.6 0.9398
    0.1 0.1 1.5 1.5 0.1938
    0.1 0.1 1.6 1.6 0.2385
    0.1 0.1 2.1 2.1 0.5313
    0.1 0.1 2.2 2.3 0.6808

     | Show Table
    DownLoad: CSV

    From Tables 25, we see that the maximum values of the correlation coefficient for the four models, FGM q-Uniform distribution, FGM q-exponential distribution, FGM q-Weibull distribution and FGM q-pareto distributionare, are different according the values of the parameters q1 and q2.

    Table 2 shows that the maximum value of the correlation coefficient ρ=0.333 when q1 and q2 with negative and positive values approach zero.

    From Table 3, we have the maximum correlation coefficient value 0.645ρ0.953 when 4q1,q210.

    Table 4 shows that the maximum correlation coefficient ρ=0.3218 when q1 and q2 take some positive values, which approaches zero. From this table, we also see that the values of ρ increases when b1 and b2 increase. In Table 5, we have the maximum correlation coefficient value 0.53ρ0.9389 for different values of q1 and q2. The values of θ1 is considered a pivotal role in increasing the values of correlation coefficient. Finally, the values of q1 and q2 play important roles to increase the maximum value of the correlation coefficient. Also, the values of q1 and q2 are not unique for the four models to increase the maximum correlation coefficient.

    In this paper, we propose the bivariate extension of the q-EWF using the FGM approach. The main topic is finding the maximum correlation coefficient for some special cases of the FGM q-EWF. For some distributions, the correlation coefficient was obtained mathematically when the rth moment of marginal distribution exists. Otherwise, the correlation coefficient is obtained numerically.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    Researchers would like to thank the Deanship of Scientific Research, Qassim University for funding publication of this project.

    The authors declare no conflict of interest.



    [1] C. A. Charalambides, Discrete q-distributions on Bernoulli trials with a geometrically varying success probability, J. Statist. Plann. Inference, 140 (2010), 2355–2383. https://doi.org/10.1016/j.jspi.2010.03.024 doi: 10.1016/j.jspi.2010.03.024
    [2] D. Cousineau, Fitting the three-parameter Weibull distribution: Review and evaluation of existing and new methods, IEEE T. Dielect. El. In., 16 (2009), 281–288. https://doi.org/10.1109/TDEI.2009.4784578 doi: 10.1109/TDEI.2009.4784578
    [3] R. Diaz, C. Ortiz, E. Pariguan, On the k-gamma q-distribution, Cent. Eur. J. Math., 8 (2010), 448–458. https://doi.org/10.2478/s11533-010-0029-0 doi: 10.2478/s11533-010-0029-0
    [4] R. Diaz, E. Pariguan, On the Gaussian q-distribution, J. Math. Anal. Appl., 358 (2009), 1–9. https://doi.org/10.1016/j.jmaa.2009.04.046 doi: 10.1016/j.jmaa.2009.04.046
    [5] H. Gauchman, Integral inequalities in q-calculus, Comput. Math. Appl., 47 (2004), 281–300. https://doi.org/10.1016/S0898-1221(04)90025-9 doi: 10.1016/S0898-1221(04)90025-9
    [6] I. S. Gradshteyn, I. M. Ryzhik, Table of integrals, series, and products (7 Eds.), Elsevier Inc., 2007.
    [7] E. J. Gumbel, Bivariate exponential distributions, J. Amer. Statist. Assoc., 299 (1960), 698–707. https://doi.org/10.1080/01621459.1960.10483368 doi: 10.1080/01621459.1960.10483368
    [8] M. Gurvich, A. Dibenedetto, S. Ranade, A new statistical distribution for characterizing the random strength of brittle materials, J. Mater. Sci., 32 (1997), 2559–2564. https://doi.org/10.1023/A:1018594215963 doi: 10.1023/A:1018594215963
    [9] P. Krugman, The self-organizing economy, Blackwell Publishers Cambridge, Massachusetts, 1996.
    [10] A. M. Mathai, A pathway to matrix-variate gamma and normal densities, Linear Algebra Appl., 396 (2005), 317–328. https://doi.org/10.1016/j.laa.2004.09.022 doi: 10.1016/j.laa.2004.09.022
    [11] D. Morgenstern, Einfache beispiele zweidimensionaler verteilungen, Mitteilungsbl. Math. Stat., 8 (1956), 234–235.
    [12] E. M. Nigm, S. Sadk, Stress-strength reliability of q-extended weibull distribution under progressive Type-II censoring, Accepted for publication in International Mathematical Forum, 2021b.
    [13] E. M. Nigm, S. W. Sadk, Stress-strength reliability for marshall-olkin q- extended Weibull family based on Type-II progressive censoring, Submitted 2021c.
    [14] S. B. Provost, A. Saboor, G. M. Cordeiro, M. Mansoor, On the q-generalized extreme value distribution, REVSTAT-Stat. J., 16 (2018), 45–70. https://doi.org/10.57805/revstat.v16i1.232 doi: 10.57805/revstat.v16i1.232
    [15] A. D. Sole, V. G. Kac, On integral representations of q-gamma and q-beta functions, Atti Accod. Naz. Lincie Cl. Fis. Mat. Nature. Rend. Linciei(9) Mat. Appl., 161 (2005), 11–29. https://doi.org/10.48550/arXiv.math/0302032 doi: 10.48550/arXiv.math/0302032
    [16] H. M. Srivastave, J. Choi, Zeta and q-zeta functions and associated series and integerals, Inc., Amsterdam, 2012.
    [17] S. Tahmasebi, A. A. Jafari, Concomitants of order statistics and record values from Morgenstern type bivariate-generalized exponential distribution, Bull. Malays. Math. Sci. Soc., 38 (2015), 1411–1423. https://doi.org/10.1007/s40840-014-0087-8 doi: 10.1007/s40840-014-0087-8
    [18] S. Tahmasebi, A. A. Jafari, M. Ahsanullah, Properties on concomitants of generalized order statistics from a bivariate Rayleigh distribution, B. Malays. Math. Sci. So., 41 (2018), 355–370. https://doi.org/10.1007/978-1-4842-4072-4_19 doi: 10.1007/978-1-4842-4072-4_19
  • This article has been cited by:

    1. M. Nagy, H. M. Barakat, M. A. Alawady, I. A. Husseiny, A. F. Alrasheedi, T. S. Taher, A. H. Mansi, M. O. Mohamed, Inference and other aspects for $ q- $Weibull distribution via generalized order statistics with applications to medical datasets, 2024, 9, 2473-6988, 8311, 10.3934/math.2024404
  • Reader Comments
  • © 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1585) PDF downloads(127) Cited by(1)

Figures and Tables

Figures(4)  /  Tables(5)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog