Processing math: 41%
Research article Special Issues

A new least squares method for estimation and prediction based on the cumulative Hazard function

  • Received: 17 March 2023 Revised: 24 June 2023 Accepted: 04 July 2023 Published: 11 July 2023
  • MSC : 60G70, 62E20, 62F10, 62G30, 62G32, 62N05

  • In this paper, the cumulative hazard function is used to solve estimation and prediction problems for generalized ordered statistics (defined in a general setup) based on any continuous distribution. The suggested method makes use of Rényi representation. The method can be used with type Ⅱ right-censored data as well as complete data. Extensive simulation experiments are implemented to assess the efficiency of the proposed procedures. Some comparisons with the maximum likelihood (ML) and ordinary weighted least squares (WLS) methods are performed. The comparisons are based on both the root mean squared error (RMSE) and Pitman's measure of closeness (PMC). Finally, two real data sets are considered to investigate the applicability of the presented methods.

    Citation: Amany E. Aly, Magdy E. El-Adll, Haroon M. Barakat, Ramy Abdelhamid Aldallal. A new least squares method for estimation and prediction based on the cumulative Hazard function[J]. AIMS Mathematics, 2023, 8(9): 21968-21992. doi: 10.3934/math.20231120

    Related Papers:

    [1] Jamalud Din, Muhammad Shabir, Nasser Aedh Alreshidi, Elsayed Tag-eldin . Optimistic multigranulation roughness of a fuzzy set based on soft binary relations over dual universes and its application. AIMS Mathematics, 2023, 8(5): 10303-10328. doi: 10.3934/math.2023522
    [2] Rukchart Prasertpong . Roughness of soft sets and fuzzy sets in semigroups based on set-valued picture hesitant fuzzy relations. AIMS Mathematics, 2022, 7(2): 2891-2928. doi: 10.3934/math.2022160
    [3] Ahmad Bin Azim, Ahmad ALoqaily, Asad Ali, Sumbal Ali, Nabil Mlaiki, Fawad Hussain . q-Spherical fuzzy rough sets and their usage in multi-attribute decision-making problems. AIMS Mathematics, 2023, 8(4): 8210-8248. doi: 10.3934/math.2023415
    [4] R. Mareay, Radwan Abu-Gdairi, M. Badr . Soft rough fuzzy sets based on covering. AIMS Mathematics, 2024, 9(5): 11180-11193. doi: 10.3934/math.2024548
    [5] Jia-Bao Liu, Rashad Ismail, Muhammad Kamran, Esmail Hassan Abdullatif Al-Sabri, Shahzaib Ashraf, Ismail Naci Cangul . An optimization strategy with SV-neutrosophic quaternion information and probabilistic hesitant fuzzy rough Einstein aggregation operator. AIMS Mathematics, 2023, 8(9): 20612-20653. doi: 10.3934/math.20231051
    [6] Amal T. Abushaaban, O. A. Embaby, Abdelfattah A. El-Atik . Modern classes of fuzzy α-covering via rough sets over two distinct finite sets. AIMS Mathematics, 2025, 10(2): 2131-2162. doi: 10.3934/math.2025100
    [7] Muhammad Naeem, Muhammad Qiyas, Lazim Abdullah, Neelam Khan, Salman Khan . Spherical fuzzy rough Hamacher aggregation operators and their application in decision making problem. AIMS Mathematics, 2023, 8(7): 17112-17141. doi: 10.3934/math.2023874
    [8] D. Jeni Seles Martina, G. Deepa . Some algebraic properties on rough neutrosophic matrix and its application to multi-criteria decision-making. AIMS Mathematics, 2023, 8(10): 24132-24152. doi: 10.3934/math.20231230
    [9] Attaullah, Shahzaib Ashraf, Noor Rehman, Asghar Khan, Muhammad Naeem, Choonkil Park . Improved VIKOR methodology based on q-rung orthopair hesitant fuzzy rough aggregation information: application in multi expert decision making. AIMS Mathematics, 2022, 7(5): 9524-9548. doi: 10.3934/math.2022530
    [10] Mona Hosny . Generalization of rough sets using maximal right neighborhood systems and ideals with medical applications. AIMS Mathematics, 2022, 7(7): 13104-13138. doi: 10.3934/math.2022724
  • In this paper, the cumulative hazard function is used to solve estimation and prediction problems for generalized ordered statistics (defined in a general setup) based on any continuous distribution. The suggested method makes use of Rényi representation. The method can be used with type Ⅱ right-censored data as well as complete data. Extensive simulation experiments are implemented to assess the efficiency of the proposed procedures. Some comparisons with the maximum likelihood (ML) and ordinary weighted least squares (WLS) methods are performed. The comparisons are based on both the root mean squared error (RMSE) and Pitman's measure of closeness (PMC). Finally, two real data sets are considered to investigate the applicability of the presented methods.



    Pawlak first proposed the rough set theory[1], which is the basis for testing the granularity of knowledge[2]. In recent years, many models related to rough sets have emerged, such as the rough set theory based on fuzzy covering[3,4,5]. The fuzzy set theory was first proposed by Zadeh[6]. Since then, theories and applications related to fuzzy sets[7] have also been widely studied, such as fuzzy soft sets[8], feature selection of fuzzy sets[9], outlier detection of fuzzy sets[10,11,12], decision application of fuzzy sets, etc.[13,14,15,16]. The relation and difference between fuzzy set theory and rough set theory is also a hot topic. An important component of this research is the roughness of the fuzzy set. Dubois et al first defined the concepts of the rough fuzzy set and fuzzy rough set[17]. The roughness measurement method of fuzzy sets proposed by Banerjee et al really makes the relationship between fuzzy sets and rough sets closer[18], and it has laid a solid foundation for subsequent researchers to explore the roughness measurement of fuzzy sets by applying fuzzy entropy[19], from the perspective of distance[20] and based on soft relation[21]. Li first mentioned the concept of disturbed fuzzy sets[22]. Liu and Chen formally described the concept of disturbed fuzzy sets[23]. Chen and Wu extended the tautology of fuzzy sets[24] to interval-valued fuzzy sets[25], intuitionistic fuzzy sets[26], and disturb fuzzy sets[27], respectively. It is found that the same result can be obtained only when the disturbation fuzzy set is generalized to the ordinary fuzzy set. Therefore, the disturbation fuzzy set shows excellent properties in the operation and has extensive research value, which is not found in any kind of fuzzy set, including the interval valued fuzzy set. Subsequently, Han et al. put forward the concept of disturbed fuzzy rough sets and the roughness measure of disturbed fuzzy sets[28]. It further enriches the theoretical basis of combining the fuzzy set and the rough set. However, there are few researches on the roughness measurement of the disturbance fuzzy set, and the application of disturbance fuzzy set roughness measurement is even less.

    Upper and lower approximations of fuzzy sets are two important aspects in the study of fuzzy rough sets theory[29]. In this paper, there are two limitations: On the one hand, it is found in the exploration that the approximation of ˜M˜N generally cannot be obtained by the approximation of ˜M and ˜N, and these properties are the result of logical forms defined by assumptions in the domain of discussion expressed in an approximate manner[30]. Therefore they bring inconvenience and difficulty to the research in many fields, including the roughness measurement of disturbed fuzzy sets. On the other hand, when data analysis, data mining, decision support system, and machine learning are carried out, the datasets are usually huge, in order to solve the inconvenience caused by too large datasets. So in this paper, first, the related concepts of the rough set, disturbation fuzzy set, and roughness measurement are introduced. Second, by introducing two new operators designed by Zhang et al. and associating them with the disturbation fuzzy set, the limitation that the execution subset is not the equality of the set and cannot be studied quantitatively is effectively solved. Finally, the roughness measure of the disturbation fuzzy set is studied quantitatively, and its boundedness is obtained. Therefore, it is expected that using the boundary of roughness measure of the disturbing fuzzy set proved in this paper can effectively avoid the computing space outside the boundary and improve the computing efficiency.

    In this section, some basic concepts related to approximate space, upper and lower approximations of fuzzy sets, the roughness measure of fuzzy sets, and disturbation fuzzy sets are given.

    Definition 2.1 (Approximate space)[1] The nonempty set U is called the discourse domain, S is the equivalence relation on U, and (U,S) is called an approximate space.

    Definition 2.2 (Upper approximation, lower approximation, and boundary field)[1] (U,S) is the known approximation space, MU, and in the approximation space y1,y2,,yk represents an equivalence class with respect to S. ˉS(M) is the upper approximation of M and S_(M) is the lower approximation of M. The boundary area BNS is represented by

    S_(M)={yi|[yi]SM},S(M)={yi|[yi]SM}, (2.1)

    while

    BNS=ˉS(M)s_(M),k=1,2,,m.

    Definition 2.3. (Upper and lower approximations of fuzzy sets) [1] In U, the upper and lower approximations of the fuzzy set M are defined as: U/S[0,1] and

    S_(M)(y)=infyYkM(x),
    ˉS(M)(y)=supyYkM(x),k=1,2,,m.

    Definition 2.4. (Roughness measure of fuzzy set)[1] (U,S) is the known approximation space, MU, and the M roughness measure in (U,S) is

    ρM=1|s_(M)||ˉS(M)|,

    where the set || represents the cardinality of .

    Yao[2] once proposed that the roughness measure of a fuzzy set can be understood as the distance between the upper approximation and the lower approximation of the fuzzy set. If M: U[0,1] is in U, M(y),yU gives y membership in M.

    Definition 2.5. (Disturbed fuzzy sets)[23] If

    ˜P:Zω,z˜P(z) (2.2)

    and

    ω={˜P(z)=(˜Pα(z),˜Pβ(z))|˜Pα(z),˜Pβ(z)[0,1]} (2.3)

    call ˜P a disturbed fuzzy set on Z, then all disturbed fuzzy sets on the discourse domain U are denoted as ˜E(U).

    Based on the concepts of upper approximation and lower approximation, this section introduces the roughness measure formula of the disturbed fuzzy set, the operation relations of upper approximation, and lower approximation, and the key properties of roughness measure of the disturbed fuzzy set.

    Definition 3.1. (Operation of disturbed fuzzy sets)[28] Let

    ω={μ=(μα,μβ)},

    the interval corresponding to (μα,μβ) is

    [max(0,μαμβ),min(1,μα+μβ)] (3.1)

    for all

    μ=(μα,μβ),ν=(να,νβ),μ,νω,

    the operation on ω is defined as

    μν=(min{μα,να},max{μβ,νβ}),μν=(max{μα,να},min{μβ,νβ}),μc=(1μα,1μβ). (3.2)

    Definition 3.2. (Relation of disturbed fuzzy sets)[28] The relationship between μ and ν is defined as

    μ=νμα=να,μβ=νβ,μνμανα,μβνβ,μ<νμα<να,μβνβ or μανα,μβ<νβ, (3.3)

    otherwise, we call it incomparable and denote by U(μ,ν).

    Obviously, when (ω,), ¯0=(0,1) and ¯1=(1,0) are the minimum and maximum elements on ω, respectively.

    Definition 3.3. (Upper and lower approximations of disturbed fuzzy sets)[28] Let μ,ν be the two given parameters,

    ˜M˜E(U),˜0<νμ˜1,

    and the (U,S) be the approximate space, defining the upper and lower approximations of a disturbed fuzzy set. The μ cut sets and ν cut sets of S_(˜M) and ˉS(˜M) are

    (S_(˜M))μ={yU|(˜M)(y)μ}, (3.4)
    (ˉS(˜M))ν={yU|ˉS(˜M)(y)ν}, (3.5)

    where, (S_(˜M))μ and (ˉS(˜M))ν can be regarded as the sets of objects with μ and ν as the minimum membership degrees in the disturbance fuzzy set ˜M.

    Definition 3.4. (Roughness of disturbed fuzzy set)[28] Let (U,S) be the approximate space,

    ˜M˜E(U),ˉ0<νμˉ1,

    then the roughness of the disturbed fuzzy set ˜M on U in accordance with parameter μ,ν is

    ˜ρ˜μ,˜ν˜M=1|S_(˜M)˜μ||ˉS(˜M)˜ν|. (3.6)

    Han et al. introduced several key properties of this roughness measure[28].

    Proposition 3.1. (Disturbation of upper and lower approximation of fuzzy sets)[28] Let μ,ν be two given parameters,

    ˜M˜E(U),0<νμ1,

    and let (S_(˜M))μ and (ˉS(˜M))ν be the μ cut sets and ν cut sets of the upper and lower approximations of the disturbed fuzzy set S_(˜M) and ˉS(˜M), where

    (ˉS(˜M˜N))ν=(ˉS(˜M))ν(ˉS(˜N))ν, (3.7)
    (S_(˜M˜N))μ=(S_(˜M))μ(S_(˜N))μ, (3.8)
    (S_(˜M))μ(S_(˜N))μ(S_(˜M˜N))μ, (3.9)
    (ˉS(˜M˜N))ν(ˉS(˜M))ν(ˉS(˜N))ν. (3.10)

    Property 3.1. For disturbed fuzzy set ˜M,˜N, there is[28]

    ˜ρμ,ν˜M˜N=1|(S_(˜M˜N))μ||(ˉS(˜M˜N))ν|=1|(S_(˜M˜N))μ||(ˉS(˜M))ν(ˉS(˜N))ν|1|(S_(˜M))μ(S_(˜N))μ||(ˉS(˜M))ν(ˉS(˜N))ν|, (3.11)
    ˜ρμ,ν˜M˜N=1|(S_(˜M˜N))μ||(ˉS(˜M˜N))ν|=1|(S_(˜M))μ(S_(˜N))μ||(ˉS(˜M˜N))ν|1|(S_(˜M))μ(S_(˜N))μ||(ˉS(˜M))ν(ˉS(˜N))ν|. (3.12)

    The pioneering study of fuzzy sets[1] derived as (3.8) and (3.9) in Proposition 3.1, which carry out the property that subsets are not equal sets, hindered the quantitative study of fuzzy sets. Because of this difficulty, Zhang et al. designed two new operators[31]. In this section, the new operator proposed by Zhang et al. is fully associated with the disturbed fuzzy set so as to effectively avoid the bad influence of this property in the roughness measurement process of the disturbed fuzzy set. The roughness measure of the disturbed fuzzy set can be studied quantitatively.

    Definition 4.1. (Determine the increment operator)[31] Let the discourse domain be U,S, the equivalence class on U, P,QU, when P is extended by Q(i.e., PQ),

    X_()():U×UU,

    defining

    X_(P)(Q)={[p]S|pH(P),hP(p)Q},

    and lP(p)Q is called the definite increment of P, where

    H(P)={hP(p)|pBNS(P)P},
    lP(p)=[p]SPandhP(p)=[p]SlP(p).

    Definition 4.2. (Uncertain decrement operator)[31] Let the discourse domain be U, S the equivalence class on U, P,QU, when P is cut by Q(i.e., PQ),

    X_()():U×UU,

    defining

    ˉX(P)(Q)={[p]S|pH(P),hP(p)Q=}

    and

    lP(p)Q,

    which is called the uncertainty decrement of P, where

    H(P)={hP(p)|pBNS(P)P},
    lP(p)=[p]SPandhP(p)=[p]SlP(p).

    Property 4.1. [31] P,QU, so

    X_(P)(Q)=X_(Q)(P), (4.1)
    ˉX(P)(Q)=ˉX(Q)(P). (4.2)

    Property 4.2. [31]

    X_(P)()=, (4.3)
    X_(P)(P)=, (4.4)
    X_(P)(¬P)=BNS(P)=ˉSPS_P. (4.5)

    Property 4.3. [31]

    ˉX(P)()=, (4.6)
    ˉX(P)(P)=, (4.7)
    ˉX(P)(¬P)=BNS(P). (4.8)

    Theorem 4.1. Let ˜M and ˜N be two disturbed fuzzy sets in the discourse domain U. Parameters μ,ν satisfy 0<νμ1, while \underline{X}˜Mμ(˜Nμ), ˉX˜Mν(˜Nν), \underline{X}˜Nμ(˜Mμ), and ˉX˜Nν(˜Mν) are, respectively, ˜Mμ, ˜Nμ determines the increment and ˜Mν, and the uncertainty of ˜Nν decreases so we can get

    (S_(˜M˜N))μ=(S_(˜M))μ(S_(˜N))μX_˜Mμ(˜Nμ)=(S_(˜M))μ(S_(˜N))μX_˜Nμ(˜Mμ), (4.9)
    (ˉS(˜M˜N))ν=(ˉS(˜M))ν(ˉS(˜N))νˉX˜Mν(˜Nν)=(ˉS(˜M))ν(ˉS(˜N))νˉX˜Nν(˜Mν). (4.10)

    Property 4.4. For disturbed fuzzy sets ˜M and ˜N,

    ˜ρμ,ν˜M˜N=1|(S_(˜M))μ(S_(˜N))μX_˜Mμ(˜Nμ)||(ˉS(˜M))ν(ˉS(˜N))ν|=1|(S_(˜M))μ(S_(˜N))μX_˜Nμ(˜Mμ)||(ˉS(˜M))ν(ˉS(˜N))ν,|, (4.11)
    ˜ρμ,ν˜M˜N=1|((˜M))μ((˜N))μ||(ˉS(˜M))ν(ˉS(˜N))νˉX˜Mν(˜Nν)|=1|((˜M))μ((˜N))μ||(ˉS(˜M))ν(ˉS(˜N))νˉX˜Nν(˜Mν)|. (4.12)

    When calculating the roughness measurement of disturbed fuzzy sets, the datasets of many programs are huge and the measurement is very complicated and cumbersome work, which requires a lot of manpower and material resources. Therefore, this section presents the boundaries of some results necessary for the roughness measurement of disturbed fuzzy sets. Understanding the boundaries of these results before operation can greatly improve work efficiency. It has very important practical significance.

    Theorem 5.1. The upper bound of the roughness measure ˜ρ˜μ,˜ν˜M˜N of the disturbed fuzzy sets ˜M and ˜N in the discourse domain U is

    ˜ρ˜μ,˜ν˜M˜N1˜ρ˜μ,˜ν˜M˜ρ˜μ,˜ν˜N2(˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N)

    with respect to parameter μ,ν, satisfying 0<νμ1.

    Proof. From (3.7) in Proposition 3.1 and the fundamental properties of sets, we can get

    ˜ρ˜μ,˜ν˜M˜N1max{|(S_(˜M))μ|,|(S_(˜N))μ|}|(ˉS(˜M))ν|+|(ˉS(˜N))ν|, (5.1)

    if

    |(S_(˜M))μ||(S_(˜N))μ|.

    Thus

    ˜ρ˜μ,˜ν˜M˜N11(|(ˉS(˜M))ν||(s_(˜M))μ|+||(ˉS(˜N))ν|||(s_(˜M))μ|), (5.2)

    so, by Definition 3.4, we get

    ˜ρ˜μ,˜ν˜M˜N1˜ρ˜μ,˜ν˜M˜ρ˜μ,˜ν˜N2(˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N). (5.3)

    Theorem 5.2. The upper bound of the roughness measure ˜ρ˜μ,˜ν˜M˜N of the disturbed fuzzy sets ˜M and ˜N in the discourse domain U is

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1+U

    with respect to parameter μ,ν, satisfying 0<νμ1 and

    U=|(s_(˜M˜N))μ||(ˉS(˜M˜N))ν|.

    Proof. From (3.8) in Proposition 3.1 and the fundamental properties of sets, we can get

    ˜ρ˜μ,˜ν˜M˜N=1|(S_(˜M))μ||(ˉS(˜M˜N))ν||(S_(˜N))μ||(ˉS(˜M˜N))ν|+|(S_(˜M˜N))μ||(ˉS(˜M˜N))ν|. (5.4)

    Also, according to (3.10) in Proposition 3.1,

    (ˉS(˜M˜N))ν(ˉS(˜M))ν, (5.5)
    (ˉS(˜M˜N))ν(ˉS(˜N))ν. (5.6)

    In other words, we have

    |(ˉS(˜M˜N))ν||(ˉS(˜M))ν|, (5.7)
    |(ˉS(˜M˜N))ν||(ˉS(˜N))ν|, (5.8)

    so we can get

    |(S_(˜M))μ||(ˉS(˜M))ν||((S_˜M))μ||(ˉS(˜M˜N))ν|, (5.9)
    |(S_(˜N))μ||(ˉS(˜N))ν||(S_(˜N))μ||(ˉS(˜M˜N))ν|. (5.10)

    Next, according to (3.9) in Proposition 3.1, it is obtained

    ˜ρ˜μ,˜ν˜M˜N1|(S_(˜M))μ||(ˉS(˜M))ν||(S_(˜N))μ||(ˉS(˜N))ν|+|(S_(˜M˜N))μ||(ˉS(˜M˜N))ν|. (5.11)

    According to Definition 3.4, we can get

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1+|(S_(˜M˜N))μ||(ˉS(˜M˜N))ν|. (5.12)

    Therefore, to sum up,

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1+U,

    when

    U=|(s_(˜M˜N))μ||(ˉS(˜M˜N))ν|.

    Remark 5.1. The bounds of Theorem 5.1 depend on roughness measures of the disturbed fuzzy sets ˜M and ˜N, and the bounds of Theorem 5.2 depend on roughness measures of the disturbed fuzzy sets ˜M and ˜N as well as (S_(˜M˜N))μ and (ˉS(˜M˜N))ν.

    Theorem 5.3. The lower bound of the disturbed fuzzy sets ˜M and ˜N in the discourse domain U for the roughness measure ˜ρ˜μ,˜ν˜M˜N with respect to parameter μ,ν is

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1+L,

    which satisfies 0<νμ1, and

    L=|X_˜Mμ(˜Nμ)|max{|(ˉS(˜M))ν|,|(ˉS(˜N))ν|}.

    Proof. From (4.11) in Property 4.4 and the fundamental properties of sets, it is obtained that

    ˜ρ˜μ,˜ν˜M˜N1|(S_(˜M))μ|+|(S_(˜N))μ|+|X_˜Mμ(˜Nμ)|max{|(ˉS(˜M))ν|,|(ˉS(˜N))ν|}, (5.13)

    We can obtain

    |(ˉS(˜M))v||(ˉS(˜N))ν| (5.14)

    and

    ˜ρ˜μ,˜ν˜M˜N1|(S_(˜M))μ|+|(S_(˜N))μ|+|X_˜Mμ(˜Nμ)||(ˉS(˜M))ν|=1|(S_(˜M))μ||(ˉS(˜M))ν||(S_(˜N))μ||(ˉS(˜M))ν||X_˜Mμ(˜Nμ)||(ˉS(˜M))ν|, (5.15)

    According to Definition 3.4 and

    |(s_(˜N))μ||(ˉS(˜M))ν||(s_(˜N))μ||(ˉS(˜N))ν|,

    we can get

    ˜ρ˜μ,˜ν˜M˜N1|(S_(˜M))μ||(ˉS(˜M))ν||(S_(˜N))μ||(ˉS(˜M))ν||X_˜Mμ(˜Nμ)||(ˉS(˜M))ν|=1(1˜ρ˜μ,˜ν˜M)(1˜ρ˜μ,˜ν˜N)|X_˜Mμ(˜Nμ)||(ˉS(˜M))ν|, (5.16)

    so

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1|X_˜Mμ(˜Nμ)||(ˉS(˜M))ν|. (5.17)

    Likewise,

    |(ˉS(˜M))v|<|(ˉS(˜N))ν|, (5.18)

    we can get

    ˜ρ˜μ,˜ν˜M˜N1|(S_(˜M))μ|+|(S_(˜N))μ|+|X_˜Mμ(˜Nμ)||(ˉS(˜N))ν|, (5.19)

    thus,

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1|X_˜Mμ(˜Nμ)||(ˉS(˜N))ν|. (5.20)

    To sum up,

    ˜ρ˜μ,˜ν˜M˜N˜ρ˜μ,˜ν˜M+˜ρ˜μ,˜ν˜N1+L,

    when

    L=|X_˜Mμ(˜Nμ)|max{|(ˉS(˜M))ν|,|(ˉS(˜N))ν|}.

    Theorem 5.4. The lower bound of the disturbed fuzzy sets ˜M and ˜N in the discourse domain U for the roughness measure ˜ρ˜μ,˜ν˜M˜N with respect to parameter μ,ν is

    ˜ρ˜μ,˜ν˜M˜N11˜ρ˜μ,˜ν˜M˜ρ˜μ,˜ν˜N+˜ρ˜μ,˜ν˜M˜ρ˜μ,˜ν˜N2˜ρ˜μ,˜ν˜M˜ρ˜μ,˜ν˜NI(1˜ρ˜μ,˜ν˜M)(1˜ρ˜μ,˜ν˜N),

    which satisfies 0<νμ1, and

    I=|(ˉS(˜M˜N))ν|+|ˉX˜Mν(˜Nν)|min{|(s_(˜M))μ|,|(s_(˜N))μ|}.

    Proof. From (4.12) in Property 4.4 and the fundamental properties of sets, it is obtained that

    ˜ρ˜μ,˜ν˜M˜N1min{|X_((˜M))μ|,|X_((˜N))μ|}|(ˉS(˜M))ν|+|(ˉS(˜N))ν||(ˉS(˜M˜N))ν||ˉX˜Mν(˜Nν)|, (5.21)

    if

    |(S_(˜M))μ||(S_(˜N))μ|, (5.22)

    we can get

    ˜ρ˜μ,˜ν˜M˜N11|(ˉS(˜M))ν||(s_(˜M))μ|+|(ˉS(˜N))ν||(s_(˜M))μ||(ˉS(˜M˜N))ν||(s_(˜M))μ||ˉX˜Mν(˜Nν)||(s_(˜M))μ|. (5.23)

    According to Definition 3.4 and

    |(ˉS(˜N))ν||(s_(˜M))μ||(ˉS(˜N))ν||((s_˜N))μ|,

    we can get

    ˜ρ˜μ,˜ν˜M˜N11|(ˉS(˜M))ν||(s_(˜M))μ|+|(ˉS(˜N))ν||(s_(˜M))μ||(ˉS(˜M˜N))ν||(s_(˜M))μ||ˉX˜Mν(˜Nν)||(s_(˜M))μ|=1111˜ρ˜μ,˜ν˜M+11˜ρ˜μ,˜ν˜N|(ˉS(˜M˜N))ν|+|ˉX˜Mν(˜Nν)||(s_(˜M))μ|. (5.24)

    Therefore, define

    I|((˜M))μ|=|(ˉS(˜M˜N))ν|+|ˉX˜Mν(˜Nν)||(s_(˜M))μ|, (5.25)

    so

    \begin{equation} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \ge 1 - \frac{{1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } }}{{2 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - {I_{\left| {{{\left(\underline{s} {\left( {\tilde M} \right)} \right)}_\mu }} \right|}}\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } } \right)\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } } \right)}}. \end{equation} (5.26)

    Likewise, for

    \begin{equation} \left| {{{\left(\underline{S} {\left( {\tilde M} \right)} \right)}_\mu }} \right| > \left| {{{\left(\underline{S} {\left( {\tilde N} \right)} \right)}_\mu }} \right|, \end{equation} (5.27)

    define

    \begin{equation} {I_{\left| {{{\left( {\left( {\tilde N} \right)} \right)}_\mu }} \right|}} = \frac{{\left| {{{\left( {\bar S\left( {\tilde M \cup \tilde N} \right)} \right)}_\nu }} \right| + \left| {{{\bar X}_{{{\tilde M}_\nu }}}\left( {{{\tilde N}_\nu }} \right)} \right|}}{{\left| {{{\left(\underline{S} {\left( {\tilde N} \right)} \right)}_\mu }} \right|}}, \end{equation} (5.28)

    so

    \begin{equation} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \ge 1 - \frac{{1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } }}{{2 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - {I_{\left| {{{\left( \underline{s}{\left( {\tilde N} \right)} \right)}_\mu }} \right|}}\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } } \right)\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } } \right)}}. \end{equation} (5.29)

    Thus, to sum up

    \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \ge 1 - \frac{{1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } + \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } }}{{2 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - {I_ \circ }\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } } \right)\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } } \right)}},

    when

    {I_ \circ } = \frac{{\left| {{{\left( {\bar S\left( {\tilde M \cup \tilde N} \right)} \right)}_\nu }} \right| + \left| {{{\bar X}_{{{\tilde M}_\nu }}}\left( {{{\tilde N}_\nu }} \right)} \right|}}{{\min \left\{ {\left| {{{\left(\underline{s} {\left( {\tilde M} \right)} \right)}_\mu }} \right|, \left| {{{\left(\underline{s} {\left( {\tilde N} \right)} \right)}_\mu }} \right|} \right\}}}.

    Remark 5.2. The lower bound of \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu, \tilde \nu } is different from the upper bound of \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu, \tilde \nu } , and the roughness measure depends not only on the disturbation fuzzy sets \tilde M and \tilde N , but also on \left| {{{\left({\bar S\left({\tilde M} \right)} \right)}_\nu }} \right|, \left| {{{\left({\bar S\left({\tilde N} \right)} \right)}_\nu }} \right|, \left| {{{\left(\underline{S} {\left({\tilde M} \right)} \right)}_\mu }} \right|, \left| {{{\left(\underline{S} {\left({\tilde N} \right)} \right)}_\mu }} \right| , and \left| {{{\bar X}_{{{\tilde M}_\nu }}}\left({{{\tilde N}_\nu }} \right)} \right| .

    Remark 5.3. In the study of the disturbed fuzzy set, it is fully understood that the roughness measure of the disturbed fuzzy set is bounded, and often roughness comparison can be made by roughly calculating the roughness measure limit of the disturbed fuzzy set, which can greatly reduce the calculation amount.

    In the previous section, it has been proved that the roughness measure of perturbed fuzzy sets is bounded, but the bound of the roughness measure of disturbed fuzzy sets can be fully applied in practical problems. Next, the superiority of the theory proposed in this paper is demonstrated more clearly through a practical application of grouping different students in a competition, as shown in Tables 15.

    Table 1.  Assessment table of students' mathematical language ability level.
    Student 1 Student 2 Student 3 Student 4 Student 5 Student 6
    \tilde M \left({0.80, 0.09} \right) \left({0.30, 0.01} \right) \left({0.66, 0.06} \right) \left({0.61, 0.23} \right) \left({0.43, 0.05} \right) \left({0.60, 0.10} \right)
    \tilde N \left({0.80, 0.09} \right) \left({0.75, 0.21} \right) \left({0.45, 0.08} \right) \left({0.70, 0.09} \right) \left({0.64, 0.03} \right) \left({0.50, 0.14} \right)

     | Show Table
    DownLoad: CSV
    Table 2.  Mathematical and verbal approximations of disturbed fuzzy sets caused by A classification.
    \rm{Student\ 1} \rm{Student\ 2}, \rm{Student\ 3}, \rm{Student\ 5} \rm{Student\ 4}, \rm{Student\ 6}
    \underline{S}\left({\tilde M} \right) \left({0.80, 0.09} \right) \left({0.30, 0.06} \right) \left({0.60, 0.23} \right)
    \underline{S}\left({\tilde N} \right) \left({0.66, 0.03} \right) \left({0.45, 0.21} \right) \left({0.50, 0.14} \right)
    \bar S\left({\tilde N} \right) \left({0.66, 0.03} \right) \left({0.75, 0.03} \right) \left({0.70, 0.09} \right)

     | Show Table
    DownLoad: CSV
    Table 3.  Approximate values of the intersection and union of mathematical and verbal disturbed fuzzy sets caused by A class classification.
    \rm{Student\ 1} \rm{Student\ 2}, \rm{Student\ 3}, \rm{Student\ 5} \rm{Student\ 4}, \rm{Student\ 6}
    \underline{S}\left({\tilde M \cup \tilde N} \right) \left({0.80, 0.03} \right) \left({0.64, 0.06} \right) \left({0.60, 0.09} \right)
    \bar S\left({\tilde M \cup \tilde N} \right) \left({0.66, 0.09} \right) \left({0.75, 0.01} \right) \left({0.70, 0.01} \right)
    \underline{S}\left({\tilde M \cap \tilde N} \right) \left({0.66, 0.09} \right) \left({0.30, 0.21} \right) \left({0.50, 0.23} \right)
    \bar S\left({\tilde M \cap \tilde N} \right) \left({0.66, 0.09} \right) \left({0.45, 0.05} \right) \left({0.61, 0.14} \right)

     | Show Table
    DownLoad: CSV
    Table 4.  Mathematical and verbal approximations of disturbed fuzzy sets caused by B classification.
    \rm{Student\ 1}, \rm{Student\ 4} \rm{Student\ 2}, \rm{Student\ 5} \rm{Student\ 3}, \rm{Student\ 6}
    \underline{S}\left({\tilde M} \right) \left({0.61, 0.23} \right) \left({0.30, 0.05} \right) \left({0.60, 0.10} \right)
    \bar S\left({\tilde M} \right) \left({0.80, 0.09} \right) \left({0.43, 0.01} \right) \left({0.66, 0.06} \right)
    \underline{S}\left({\tilde N} \right) \left({0.66, 0.09} \right) \left({0.64, 0.21} \right) \left({0.45, 0.14} \right)
    \bar S\left({\tilde N} \right) \left({0.70, 0.03} \right) \left({0.75, 0.03} \right) \left({0.50, 0.08} \right)

     | Show Table
    DownLoad: CSV
    Table 5.  Approximate values of the intersection and union of mathematical and verbal disturbed fuzzy sets caused by B class classification.
    \rm{Student\ 1}, \rm{Student\ 4} \rm{Student\ 2}, \rm{Student\ 5} \rm{Student\ 3}, \rm{Student\ 6}
    \underline{S}\left({\tilde M \cup \tilde N} \right) \left({0.70, 0.09} \right) \left({0.64, 0.03} \right) \left({0.60, 0.10} \right)
    \bar S\left({\tilde M \cup \tilde N} \right) \left({0.80, 0.03} \right) \left({0.75, 0.01} \right) \left({0.66, 0.06} \right)
    \underline{S}\left({\tilde M \cap \tilde N} \right) \left({0.61, 0.23} \right) \left({0.30, 0.21} \right) \left({0.45, 0.14} \right)
    \bar S\left({\tilde M \cap \tilde N} \right) \left({0.66, 0.09} \right) \left({0.43, 0.05} \right) \left({0.50, 0.08} \right)

     | Show Table
    DownLoad: CSV

    Example 6.1. Due to receiving the notice that our province will soon hold a student learning competition to test the learning ability of two subjects of mathematics and Chinese, the school will send 6 students to participate in the competition. It is known that each student's ability level assessment of mathematics and Chinese constitutes a disturbance fuzzy set. The school will formulate two combinations, respectively,

    A:\left\{ {\left\{ {{\rm{student}}1} \right\}, \left\{ {{\rm{student}}2, {\rm{student}}3, {\rm{student}}5} \right\}, \left\{ {{\rm{student}}4, {\rm{student}}6} \right\}} \right\},
    B:\left\{ {\left\{ {{\rm{student}}1, {\rm{student}}4} \right\}, \left\{ {{\rm{student}}2, {\rm{student}}5} \right\}, \left\{ {{\rm{student}}3, {\rm{student}}6} \right\}} \right\}.

    If you want to know which combination is more likely to win, set parameter

    \left( {0.00, 0.00} \right) < \mu = \nu \le \left( {0.60, 0.10} \right),

    (in real life, people usually think that 60 is a passing grade on a 100-point scale, and the parameter selection of different practical questions will be different). Table 1 is the assessment table of students' mathematical and language ability levels. The mathematics of disturbed fuzzy sets and the language of disturbed fuzzy sets are represented by \tilde M and \tilde N , respectively.

    So, according to Definition 3.4 and Tables 2 and 4,

    \begin{equation} \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } (A) = 1 - \frac{{\left|\underline{S} {{{\left( {\tilde M} \right)}_{\tilde \mu }}} \right|}}{{\left| {\bar S{{\left( {\tilde M} \right)}_{\tilde \nu }}} \right|}} = 1 - \frac{1}{6} = \frac{5}{6}, \qquad\mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } (A) = 1 - \frac{{\left| \underline{S}{{{\left( {\tilde N} \right)}_{\tilde \mu }}} \right|}}{{\left| {\bar S{{\left( {\tilde N} \right)}_{\tilde \nu }}} \right|}} = 1 - \frac{1}{6} = \frac{5}{6}, \end{equation} (6.1)
    \begin{equation} \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } (B) = 1 - \frac{{\left| \underline{S}{{{\left( {\tilde M} \right)}_{\tilde \mu }}} \right|}}{{\left| {\bar S{{\left( {\tilde M} \right)}_{\tilde \nu }}} \right|}} = 1 - \frac{2}{4} = \frac{1}{2}, \qquad\mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } (B) = 1 - \frac{{\left| \underline{S}{{{\left( {\tilde N} \right)}_{\tilde \mu }}} \right|}}{{\left| {\bar S{{\left( {\tilde N} \right)}_{\tilde \nu }}} \right|}} = 1 - \frac{2}{4} = \frac{1}{2}. \end{equation} (6.2)

    From Theorems 5.1–5.4, it follows

    \begin{align} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( A \right)& \le \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } + \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - 1 + {U^ \circ } = \frac{{20}}{3}, \end{align} (6.3)
    \begin{align} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( A \right) &\ge 1 - \frac{{1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } + \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } }}{{2 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - {I_ \circ }\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } } \right)\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } } \right)}} = \frac{{17}}{{18}}, \end{align} (6.4)
    \begin{align} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( B \right) &\le \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } + \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - 1 + {U^ \circ } = 3, \end{align} (6.5)
    \begin{align} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( B \right)& \ge 1 - \frac{{1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } + \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } }}{{2 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } - {I_ \circ }\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde M}^{\tilde \mu , \tilde \nu } } \right)\left( {1 - \mathop {\tilde \rho }\nolimits_{\tilde N}^{\tilde \mu , \tilde \nu } } \right)}} = \frac{6}{7}, \end{align} (6.6)

    and to sum up,

    \frac{{17}}{{18}}\mathop { \le \tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( A \right) \le \frac{{20}}{3}, \quad\mathop {3 \le \tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\tilde \mu , \tilde \nu } \left( B \right) \le \frac{6}{7}.

    Obviously, the roughness of B classification is smaller.

    If the traditional disturbation fuzzy set roughness measure calculation method is

    \begin{equation} \mathop {\tilde \rho }\nolimits_{\tilde M \cap \tilde N}^{\mu , \nu } = 1 - \frac{{\left| {{{\left(\underline{S} {\left( {\tilde M \cap \tilde N} \right)} \right)}_\mu }} \right|}}{{\left| {{{\left( {\bar S\left( {\tilde M \cap \tilde N} \right)} \right)}_\nu }} \right|}} = 1 - \frac{{\left| {{{\left(\underline{S} {\left( {\tilde M} \right)} \right)}_\mu } \cap {{\left(\underline{S} {\left( {\tilde N} \right)} \right)}_\mu }} \right|}}{{\left| {{{\left( {\bar S\left( {\tilde M \cap \tilde N} \right)} \right)}_\nu }} \right|}}, \end{equation} (7.1)

    we need to calculate the number of equivalence classes after the intersection of {\left(\underline{S}{\left({\tilde M} \right)} \right)_\mu } and {\left(\underline{S}{\left({\tilde N} \right)} \right)_\mu } .

    In Tables 3 and 5, the approximate values of the mathematical and verbal intersection of the disturbed fuzzy set caused by A classification and B classification are listed, respectively. It can be seen that the traditional method is more complicated to calculate. However, it can be seen from the example that using the method proposed in this paper to avoid complex calculation can effectively improve the work efficiency. This paper only lists 2 classification methods for 6 students. In practical problems, there may be tens of thousands of students' classification methods, etc. Therefore, when the sample size is large, the roughness measurement boundary of the disturbation fuzzy set proposed in this paper will greatly reduce the workload in operation. In practical problems with large datasets, such as when we need to do data mining, bioinformatics, cybersecurity, natural language processing, etc., the sample size is often huge. Therefore, it is usually better to determine the range of roughness first and then calculate in a small range.

    First, this work effectively solves the problem that the execution subsets are not equal sets, which hindrances the quantitative study of disturbed fuzzy sets.

    Second, through quantitative research, the new properties of the disturbation fuzzy set operation and the boundary of the roughness of the disturbation fuzzy set are established effectively, which can effectively reduce the workload in the operation when the actual data capacity is huge.

    This paper proposes and proves that the roughness measure of the disturbed fuzzy set is bounded. In practical application, a full understanding of the roughness measure boundary of the disturbed fuzzy set can effectively avoid unnecessary computing space and greatly improve work efficiency. However, the roughness measurement of disturbed fuzzy sets depends on the choice of parameter \mu, \nu . The roughness measurement of disturbed fuzzy sets without parameters will be further explored in future work.

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

    This work was supported by the National Natural Science Foundation of China (No. 61603055).

    Li Li: responsible for the planning, design and implementation of the research, providing financial and technical support. Hangyu Shi: designed research methods, processed and analyzed data, performed theorem proving, and wrote the first draft of the paper. Xiaona Liu: assisted in paper analysis and verification, and participated in paper revision. Jingjun Shi: provide partial data and coordinate the study as a whole.

    The authors declare no conflicts of interest.



    [1] N. Balakrishnan, E. Cramer, The art of progressive censoring: Applications to reliability and quality, Boston, MA: Birkhäuser, 2014. https://doi.org/10.1007/978-0-8176-4807-7
    [2] G. Casella, R. L. Berger, Statistical inference, Duxbury Press, Pacific Grove, CA., 2002.
    [3] E. L. Lehmann, G. Casella, Theory of point estimation, 2 Eds., New York: Springer-Verlag, 1998.
    [4] M. Ahsanullah, Linear prediction of record values for the two parameter exponential distribution, Ann. Inst. Stat. Math., 32 (1980), 363–368. https://doi.org/10.1007/BF02480340 doi: 10.1007/BF02480340
    [5] E. K. AL-Hussaini, Predicting observables from a general class of distributions, J. Stat. Plan. Infer., 79 (1999), 79–91. https://doi.org/10.1016/S0378-3758(98)00228-6 doi: 10.1016/S0378-3758(98)00228-6
    [6] E. K. AL-Hussaini, A. B. Ahmad, On Bayesian predictive distributions of generalized order statistics, Metrika, 57 (2003), 165–176. https://doi.org/10.1007/s001840200207 doi: 10.1007/s001840200207
    [7] A. E. Aly, Prediction of the exponential fractional upper record values, Math. Slovaca, 72 (2022), 491–506. http://dx.doi.org/10.1515/ms-2022-0032 doi: 10.1515/ms-2022-0032
    [8] A. E. Aly, Predictive inference of dual generalized order statistics from the inverse Weibull distribution, Stat. Pap., 64 (2023), 139–160. http://dx.doi.org/10.1007/s00362-022-01312-0 doi: 10.1007/s00362-022-01312-0
    [9] M. Amleh, M. Raqab, Prediction of censored Weibull lifetimes in a simple step-stress plan with Khamis-Higgins model, Stat. Optim. Inf. Comput., 10 (2022), 658–677. http://dx.doi.org/ 10.19139/soic-2310-5070-1069 doi: 10.19139/soic-2310-5070-1069
    [10] H. M. Barakat, O. M. Khaled, H. A. Ghanem, New method for prediction of future order statistics, Qual. Technol. Quant M., 18 (2021), 101–116.
    [11] H. M. Barakat, M. E. El-Adll, A. E. Aly, Two-sample nonparametric prediction intervals based on random number of generalized order statistics, Commun. Stat. Theor. M., 50 (2021), 4571–4586. http://dx.doi.org/10.1080/03610926.2020.1719421. doi: 10.1080/03610926.2020.1719421
    [12] M. E. El-Adll, H. M. Barakat, A. E. Aly, Asymptotic prediction for future observations of a random sample of unknown continuous distribution, Complexity, 2022 (2022), 4571–4586. http://dx.doi.org/10.1155/2022/4073799 doi: 10.1155/2022/4073799
    [13] K. S. Kaminsky, L. S. Rhodin, Maximum likelihood prediction, Ann. Inst. Stat. Math., 37 (1985), 507–517. https://doi.org/10.1007/BF02481119 doi: 10.1007/BF02481119
    [14] J. F. Lawless, A prediction problem concerning samples from the exponential distribution with applications in life testing, Technometrics, 13 (1971), 725–730. http://dx.doi.org/10.2307/1266949 doi: 10.2307/1266949
    [15] G. S. Lingappaiah, Prediction in exponential life testing, Can. J. Stat., 1 (1973), 113–117. http://dx.doi.org/10.2307/3314650 doi: 10.2307/3314650
    [16] Z. M. Raqab, Optimal prediction-intervals for the exponential distribution based on generalized order statistics, IEEE Trans. Reliab., 50 (2001), 112–115. https://doi.org/10.1109/24.935025 doi: 10.1109/24.935025
    [17] J. Aitcheson, I. Dunsmore, Statistical prediction analysis, Cambridge University Press, Cambridge, 1975. https://doi.org/10.1017/CBO9780511569647
    [18] H. A. David, H. N. Nagaraja, Order statistics, 3 Eds., Wiley, NJ, 2003. https://doi.org/10.1002/0471722162
    [19] S. Geisser, Predictive inference: An introduction, Chapman and Hall: London, 1993.
    [20] K. S. Kaminsky, P. I. Nelson, Prediction intervals, In: Handbook of Statistics, Balakrishnan, N. and C. R. Rao (Eds.), Amesterdam, North Holland, 50 (1998), 431–450. https://doi.org/10.7788/bue-1997-jg33
    [21] H. M. Barakat, M. E. El-Adll, M. E. Sobh, Bootstrapping m-generalized order statistics with variable rank, AIMS Math., 7 (2022), 13704–13732. http://dx.doi.org/10.3934/math.2022755 doi: 10.3934/math.2022755
    [22] U. Kamps, A concept of generalized order statistics, Teubner, Stuttgart, 1995. https://doi.org/10.1007/978-3-663-09196-7
    [23] U. Kamps, E. Cramer, On distributions of generalized order statistics, In: Handbook of Statistics, Balakrishnan, N. and C. R. Rao (Eds.), Amesterdam, North Holland, 35 (2001), 269–280. http://dx.doi.org/10.1080/02331880108802736
    [24] J. Swain, S. Venkatraman, J. Wilson, Least squares estimation of distribution function in Johnson's translation system, J. Stat. Comput. Simul., 29 (1988), 271–297. https://doi.org/10.1080/00949658808811068 doi: 10.1080/00949658808811068
    [25] R. D. Gupta, D. Kundu, Generalized exponential distribution: Different method of etimations, J. Stat. Comput. Simul., 69 (2001), 315–337. https://doi.org/10.1080/00949650108812098 doi: 10.1080/00949650108812098
    [26] D. Kundu, M. Z. Raqab, Generalized Rayleigh distribution: Different methods of estimation, Comput. Stat. Data Anal., 49 (2005), 187–200. https://doi.org/10.1016/j.csda.2004.05.008 doi: 10.1016/j.csda.2004.05.008
    [27] M. E. El-Adll, A. E. Aly, Prediction intervals for future observations of pareto distribution based on generalized order statistics, J. Appl. Statist. Sci., 22 (2016), 111–125.
    [28] H. M. Barakat, E. M. Nigm, M. E. El-Adll, M. Yusuf, Prediction of future exponential lifetime based on random number of generalized order statistics under a general set-up, Stat. Pap., 59 (2018), 605–631. http://dx.doi.org/10.1007/s00362-016-0779-2 doi: 10.1007/s00362-016-0779-2
    [29] A. Rényi, On the theory of order statistics, Acta Math. Hungarica, 4 (1953), 191–231. https://doi.org/10.1007/BF02127580 doi: 10.1007/BF02127580
    [30] E. J. G. Pitman, J. Wishart, The closest estimates of statistical parameters, Math. Proc. Cambridge Philos. Soc., 59 (1937), 212–222. https://doi.org/10.1017/S0305004100019563 doi: 10.1017/S0305004100019563
    [31] J. P. Keating, R. L. Mason, P. K. Sen, Pitman's measure of closeness: A comparison of statistical estimators, SIAM, Philadelphia: Society for Industrial and Applied Mathematics, 1993.
    [32] N. Balakrishnan, K. F. Davies, J. P. Keating, R. L. Mason, Pitman closeness of best linear unbiased and invariant predictors for exponential distribution in one- and two-sample situations, Commun. Stat.-Theor. M., 41 (2012), 1–15. https://doi.org/10.1080/03610920903537301 doi: 10.1080/03610920903537301
    [33] H. N. Nagaraja, Comparison of estimators and predictors from two-parameter exponential distribution, Sankhya Ser B., 48 (1986), 10–18.
    [34] Z. M. Raqab, A. L. Alkhalfan, N. Balakrishnan, Pitman comparisons of predictors of censored observations from progressively censored samples for exponential distribution, J. Stat. Comput. Simul., 86 (2016), 1539–1558. https://doi.org/10.1080/00949655.2015.1071820 doi: 10.1080/00949655.2015.1071820
    [35] M. E. El-Adll, Inference for the two-parameter exponential distribution with generalized order statistics, Math. Popul. Stud., 28 (2021), 1–23. http://dx.doi.org/10.1080/08898480.2019.1681187 doi: 10.1080/08898480.2019.1681187
    [36] A. A. Al-Babtain, M. K. Shakhatreh, M. Nassar, A. Z. Afify, A new modified Kies family: Properties, estimation under complete and type-Ⅱ censored samples, and engineering applications, Mathematics, 8 (2020), 1345. http://dx.doi.org/10.3390/math8081345 doi: 10.3390/math8081345
    [37] A. M. Abd El-Raheem, E. M. Almetwally, M. S. Mohamed, E. H. Hafez, Accelerated life tests for modified Kies exponential lifetime distribution: Binomial removal, transformers turn insulation application and numerical results, AIMS Math., 6 (2021), 5222–5255. https://doi.org/10.3934/math.2021310 doi: 10.3934/math.2021310
    [38] H. M. Barakat, M. E. El-Adll, A. E. Aly, Prediction of future generalized order statistics based on two-parameter exponential distribution for large samples, Qual. Technol. Quant. Manag., 19 (2022), 259–275. http://dx.doi.org/10.1080/16843703.2022.2034261 doi: 10.1080/16843703.2022.2034261
    [39] D. N. P. Murthy, M. Xie, R. Jiang, Weibull models, Wiley: Hoboken, NJ, USA, 2004.
    [40] D. G. Hoel, A representation of mortality data by competing risks, Biometrics, 28 (1972), 475–488. https://doi.org/10.2307/2556161 doi: 10.2307/2556161
    [41] W. S. Abu El Azm, R. A. Aldallal, H. M. Aljohani, S. G. Nassr, Estimations of competing lifetime data from inverse Weibull distribution under adaptive progressively hybrid censored, Math. Biosci. Eng., 19 (2022), 6252–6275. https://doi.org/10.3934/mbe.2022292 doi: 10.3934/mbe.2022292
  • This article has been cited by:

    1. Li Li, Jin Yang, Disturbing Fuzzy Multi-Attribute Decision-Making Method with If Weight Information Is Disturbing Fuzzy Number, 2024, 12, 2227-7390, 1225, 10.3390/math12081225
  • 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(1791) PDF downloads(95) Cited by(0)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog