Abstract
In this article we propose and study a new family of distributions which is defined by using the genesis of the truncated Poisson distribution and the exponentiated generalized-G distribution. Some mathematical properties of the new family including ordinary and incomplete moments, quantile and generating functions, mean deviations, order statistics and their moments, reliability and Shannon entropy are derived. Estimation of the parameters using the method of maximum likelihood is discussed. Although this generalization technique can be used to generalize many other distributions, in this study we present only two special models. The importance and flexibility of the new family is exemplified using real world data.
1 Introduction
In the last decades, several generalized distributions have been proposed based on different modification methods. These modification methods require the addition of one or more parameters to base model which could provide better adaptability in the modeling of real lifetime data. Modern computing technology has made many of these techniques accessible even if analytical solutions are very complicated. Several continuous univariate-G families have recently appeared. Some notable family includes Marshall–Olkin-G family by Marshall and Olkin [22], exponentiated-G class by R. C. Gupta, P. L. Gupta and R. D. Gupta [17], transmuted exponentiated generalized-G family by Yousof, Afify, Alizadeh, Butt, Hamedani and Ali [36], transmuted geometric-G by Afify, Alizadeh, Yousof, Aryal and Ahmad [1], Kumaraswamy transmuted-G by Afify, Cordeiro, Yousof, Alzaatreh and Nofal [2], Burr X-G by Yousof, Afify, Hamedani and Aryal [37], the odd Lindley-G family of distributions by Silva, Percontini, de Brito, Ramos, Venancio and Cordeiro [33], exponentiated transmuted-G family by Merovci, Alizadeh, Yousof and Hamedani [23], the odd-Burr generalized family by Alizadeh, Cordeiro, Nascimento, Lima and Ortega [5], the transmuted Weibull-G family by Alizadeh, Rasekhi, Yousof and Hamedani [6], the type I half-logistic family by Cordeiro, Alizadeh and Diniz Marinho [11], the complementary generalized transmuted Poisson family by Alizadeh, Yousof, Afify, Cordeiro and Mansoor [7], the Zografos–Balakrishnan odd log-logistic family of distributions by Cordeiro, Alizadeh, Ortega and Serrano [12], logistic-X by Tahir, Cordeiro, Alzaatreh, Mansoor and Zubair [34], a new Weibull-G by Tahir, Zubair, Mansoor, Cordeiro, Alizadeh and Hamedani [35], the generalized odd log-logistic family by Cordeiro, Alizadeh, Ozel, Hosseini, Ortega and Altun [13], the beta odd log-logistic generalized family of distributions by Cordeiro, Alizadeh, Tahir, Mansoor, Bourguignon and Hamedani [14], beta transmuted-H by Afify, Yousof and Nadarajah [3], generalized transmuted-G by Nofal, Afify, Yousof and Cordeiro [30] and beta Weibull-G family by Yousof, Rasekhi, Afify, Ghosh, Alizadeh and Hamedani [38] among others.
In this paper we propose and study a generalized family of distribution using the genesis of Poisson distribution with the following motivation. Suppose that a system has N subsystems functioning independently at a given time where N has zero truncated Poisson (ZTP) distribution with parameter λ. It is the conditional probability distribution of a Poisson-distributed random variable, given that the value of the random variable is not zero. The probability mass function (pmf) of N is given by
Note that for ZTP variable the expected value and variance are respectively given by
and
Suppose that the failure time of each subsystem has the exponentiated
Generalized-G (“
and
respectively, where
Therefore, the unconditional cdf of X, as described in [31], can be expressed as
The cdf in (1) is called the exponentiated generalized G Poisson (“EGGP”) family of distributions. The corresponding pdf is
For
Using the power series expansion of
Using the series expansion
where
and
This is the Exp-G pdf with power parameter
where
The properties of Exp-G distributions have been studied by many authors in recent years, see Mudholkar and Srivastava [25] and Mudholkar, Srivastava and Freimer [26] for exponentiated Weibull (EW) distributions, R. C. Gupta, P. L. Gupta and R. D. Gupta [17] for exponentiated Pareto distributions, Gupta and Kundu [18] for exponentiated exponential distributions, Nadarajah and Kotz [29] for the exponentiated-type distributions, Nadarajah [27] for exponentiated Gumbel distributions, Shirke and Kakade [32] for exponentiated log-normal distributions and Nadarajah and Gupta [28] for exponentiated gamma distributions (EGa), among others.
The rest of the paper is outlined as follows. In Section 2 we provide the formulation of EGGP models for two special distributions. Mathematical properties of the EGGP model are discussed in Section 3. In Section 4 we discuss stress-strength models. The order statistics is discussed in Section 5. Parameter estimation procedures using method of maximum likelihood are presented in Section 6. Section 7 provides the application of the two generalized distributions to model real world data. Some concluding remarks are given in Section 8.
2 Special Models
The formulation provided in Section 1 can be used to generalize any classical probability distribution. For illustration purpose we will generalize the following two popular and versatile distributions, namely: the Weibull (W) distribution and the Pareto (Pa) distribution. The parameters of these models are positive real numbers. The pdf and cdf of these distributions are provided in Table 1.
Model | Pdf: | Cdf: | Support |
W | |||
Pa |
2.1 The EGWP Distribution
The cdf and pdf of the EG-Weibull Poisson (EGWP) distribution are given, respectively, by
and
Plots of the pdf and cdf of the EGWP distribution are displayed in Figure 1 for some parameter values. As we shall see from the graphs EGWP distribution is more flexible compare to classical Weibull distribution.
2.2 The EGPaP Distribution
The cdf and pdf of the EG-Pareto Poisson(EGPaP) distribution (for
and
Plots of the pdf and cdf of the EGPaP distribution are displayed in Figure 2 for some parameter values. As we shall see from the graphs EGPaP distribution is more flexible compare to the Pareto distribution.
3 Mathematical Properties
In this section we provide some structural and mathematical properties of the EGGP distribution including the quantile function, moments, entropy measure, residual and revered residual life.
3.1 Quantile Function
The quantile function of a distribution is the real solution of
we have
Therefore
We can use the inversion method to simulate random numbers from a given distribution. For example, we can simulate random numbers X from EGWP distribution by
which implies
where U has uniform distribution on
3.2 General Properties
The rth ordinary moment of X is given by
Henceforth,
which can be computed numerically in terms
of the baseline quantile function (qf)
Setting
and
where
The cumulants (
where
The main application of the first incomplete moment refers
to the Bonferroni and Lorenz curves. These curves are very useful in
economics, reliability, demography, insurance and medicine. The answers to
many important questions in economics require more than just knowing the
mean of the distribution, but its shape as well. This is obvious not only
in the study of econometrics but in other areas as well. The sth
incomplete moments, say
The first incomplete moment of the EGGP family,
and
respectively, where
Now, we provide two ways to determine
where
is the first
incomplete moment of the Exp-G distribution. A second general formula for
where
can be computed numerically. These
equations for
and
3.3 Probability Weighted Moments
The PWMs are expectations of certain functions of a random variable and they
can be defined for any random variable whose ordinary moments exist. The PWM
method can generally be used for estimating parameters of a distribution
whose inverse form cannot be expressed explicitly. The
Using equations (2) and (3), we can write
where
Then the
3.4 Entropy Measures
The Rényi entropy of a random variable X represents a measure of variation of the uncertainty. The Rényi entropy is defined by
Using the power series expansion, the pdf in (2) can be expressed as
where
Therefore, the Rényi entropy of the EGGP family is given by
The q-entropy, say
where
The Shannon entropy of a random variable X, say
It is the special case of the Rényi entropy when
3.5 Residual Life and Reversed Residual Life
The nth moment of the residual life, say
Therefore
where
Then the nth moment of the reversed residual life of X becomes
where
and
4 Stress-Strength Models
The stress-strength model is the most widely used approach for reliability estimation. This model is used in many applications of physics and engineering, such as strength failure and system collapse. In stress-strength modeling,
is a measure of reliability of the system when it is subjected to random
stress
Then
where
5 Order Statistics
Let
where
where
Moreover, the pdf of
therefore, the density function of the EGGP order statistics is a mixture of EG
densities. Based on the last equation, we note that the properties of
For the EGWP model we have
The L-moments are analogous to the ordinary moments but can be estimated by linear combinations of order statistics. They exist whenever the mean of the distribution exists, even though some higher moments may not exist, and are relatively robust to the effects of outliers. Based upon the moments in equation (10), we can derive explicit expressions for the L-moments of X. They are linear functions of expected order statistics defined by
The first four L-moments are given by
One can simply obtain the L-moments
6 Estimation
Let
where
are given by
and (for
where
Setting the nonlinear system of equations
7 Applications
In many statistical applications, the interest is centered on estimating the
parameters and evaluate the goodness-of-fit of the model to analyze the data
on hand. In this section, we provide the effectiveness of the EGGP distribution by means of modeling two different data sets choosing two special models discussed in Section 2. These data sets have been used by several authors to show the applicability of other competing models. We also provide a formative evaluation
of the goodness-of-fit of the models and make comparisons with other distributions. The measures of goodness-of-fit,
including the Akaike information criterion (AIC), Bayesian information
criterion (BIC), Anderson–Darling (
Example 1: Cancer Patient Data.
This data set describes the remission times (in months) of a random sample
of 128 bladder cancer patients studied by Lee and Wang [21]. For these
data, we compare the fit of the EGWP with the other five parameter distributions which has
been generalized using the Weibull genesis. We compare the fits of the EGWP with the generalized transmuted-W (GTW) distribution
(Nofal, Afify, Yousof and Cordeiro [30]), the McDonald Weibull (McW) distribution (Cordeiro, Hashimoto and Ortega [15]), the modified beta
Weibull (MBW) distribution (Khan [19]) and the transmuted additive Weibull (TAW) distribution (Elbatal and
Aryal [16]) with the corresponding densities given by (for
The parameters of the above densities are all positive real numbers except
Goodness-of-fit criteria | ||||||
Model | AIC | CAIC | HQIC | BIC | ||
EGWP | 829.448 | 829.939 | 835.242 | 843.708 | 0.0227 | 0.1505 |
GTW | 831.347 | 831.839 | 837.141 | 845.607 | 0.0469 | 0.3058 |
McW | 831.680 | 832.172 | 837.474 | 845.94 | 0.0504 | 0.3299 |
MBW | 838.027 | 838.519 | 843.821 | 852.288 | 0.1068 | 0.7207 |
TAW | 838.478 | 838.97 | 844.272 | 852.739 | 0.1129 | 0.7033 |
Estimates | |||||
Model | |||||
EGWP | 0.4202 | 1.5857 | 0.1540 | 0.9237 | 3.7776 |
(2.3728) | (0.7778) | (0.9355) | (0.3111) | (2.4207) | |
GTW | 2.7965 | 0.0128 | 0.2991 | 0.6542 | 0.002 |
(1.117) | (7.214) | (0.151) | (0.121) | (1.769) | |
McW | 4.0633 | 2.6036 | 0.1192 | 0.5582 | 0.0393 |
(2.111) | (2.452) | (0.109) | (0.178) | (0.202) | |
MBW | 57.4167 | 19.3859 | 10.1502 | 0.1632 | 2.0043 |
(37.317) | (13.490) | (22.437) | (0.044) | (0.789) | |
TAW | 0.00003 | 1.0065 | 0.1139 | 0.9722 | |
(0.0061) | (0.035) | (0.032) | (0.125) | (0.280) |
Example 2: Flood Data.
This data set describes the exceedances of flood peaks (in
The parameters of the above densities are all positive real numbers except
Model | |||||
EGPaP | 6.5163 | 4.9880 | 20.4148 | 0.0264 | 0.1 |
(2.2125) | (0.8487) | (8.9005) | (0.0088) | – | |
KwP | 2.8553 | 85.8468 | – | 0.0528 | 0.1 |
(0.3371) | (60.4213) | – | (0.0185) | – | |
BP | 3.1473 | 85.7508 | – | 0.0088 | 0.1 |
(0.4993) | (0.0001) | – | (0.0015) | – | |
TP | 1 | 1 | 0.3490 | 0.1 | |
– | – | (0.089) | (0.072) | – | |
EP | 2.8797 | 1 | – | 0.4241 | 0.1 |
(0.4911) | – | – | (0.0463) | – | |
P | 1 | 1 | – | 0.2438 | 0.1 |
– | – | – | (0.0287) | – |
Statistics | ||||||
Model | AIC | CAIC | BIC | HQIC | KS | |
EGPaP | 255.131 | 520.262 | 521.171 | 531.645 | 524.794 | 0.1428 |
KwP | 271.200 | 548.400 | 548.753 | 555.230 | 551.119 | 0.1700 |
BP | 283.700 | 573.400 | 573.753 | 580.230 | 576.119 | 0.1747 |
TP | 286.201 | 576.402 | 576.575 | 580.954 | 578.214 | 0.2870 |
EP | 287.300 | 578.600 | 578.774 | 583.153 | 580.413 | 0.1987 |
P | 303.100 | 608.200 | 608.257 | 610.477 | 609.106 | 0.3324 |
Fitted pdf, cdf and QQ-plots for both data are provided in Figure 3. It can be observed that the EGWP distribution is appropriate to model the cancer patient data and the EGPaP distribution is appropriate to model the flood peak exceedance data.
8 Conclusions
In this study, we have introduced the so-called exponentiated generalized G-Poisson family of distribution. Some mathematical properties of the new family including ordinary and incomplete moments, quantile and generating functions, mean deviations, order statistics and their moments, reliability and Shannon entropy are derived. Although this generalization technique can be used to generalize many other distributions, for illustration purposes we have chosen the Weibull distribution and the Pareto distribution as base distributions. The importance and flexibility of the new family are illustrated by means of two different examples, one for each generalized family. We hope that this study will serve as a reference and help to advance future research in the subject area.
A Appendix
The elements of the observed matrix
and
where
and
Acknowledgements
The authors are grateful to the editor and anonymous reviewer for their constructive comments and valuable suggestions which certainly improved the presentation and quality of the paper.
References
[1] A. Z. Afify, M. Alizadeh, H. M. Yousof, G. Aryal and M. Ahmad, The transmuted geometric-G family of distributions: Theory and applications, Pak. J. Statist. 32 (2016), 139–160. Search in Google Scholar
[2] A. Z. Afify, G. M. Cordeiro, H. M. Yousof, A. Alzaatreh and Z. M. Nofal, The Kumaraswamy transmuted-G family of distributions: Properties and applications, J. Data Sci. 14 (2016), 245–270. 10.6339/JDS.201604_14(2).0004Search in Google Scholar
[3] A. Z. Afify, H. M. Yousof and S. Nadarajah, The beta transmuted-H family for lifetime data, Stat. Interface 10 (2017), 505–520. 10.4310/SII.2017.v10.n3.a13Search in Google Scholar
[4] A. Akinsete, F. Famoye and C. Lee, The beta-Pareto distribution, Statistics 42 (2008), 547–563. 10.1080/02331880801983876Search in Google Scholar
[5] M. Alizadeh, G. M. Cordeiro, A. D. C. Nascimento, M. D. S. Lima and E. M. M. Ortega, Odd-Burr generalized family of distributions with some applications, J. Stat. Comput. Simul. 83 (2016), 326–339. 10.1080/00949655.2016.1209200Search in Google Scholar
[6] M. Alizadeh, M. Rasekhi, H. M. Yousof and G. G. Hamedani, The transmuted Weibull-G family of distributions, Hacet. J. Math. Stat. (2017), 10.15672/HJMS.2017.440. 10.15672/HJMS.2017.440Search in Google Scholar
[7] M. Alizadeh, H. M. Yousof, A. Z. Afify, G. M. Cordeiro and M. Mansoor, The complementary generalized transmuted Poisson-G family, Austrian J. Stat., to appear. 10.17713/ajs.v47i4.577Search in Google Scholar
[8] M. Bourguignon, M. B. Silva, L. M. Zea and G. M. Cordeiro, The Kumaraswamy Pareto distribution, J. Stat. Theory Appl. 12 (2013), 129–144. 10.2991/jsta.2013.12.2.1Search in Google Scholar
[9] G. Chen and N. A. Balakrishnan, A general purpose approximate goodness-of-fit test, J. Quality Technol. 27 (1995), 154–161. 10.1080/00224065.1995.11979578Search in Google Scholar
[10] V. Choulakian and M. A. Stephenes, Goodness-of-fit for the generalized Pareto distribution, Technometrics 43 (2001), 478–484. 10.1198/00401700152672573Search in Google Scholar
[11] G. M. Cordeiro, M. Alizadeh and P. R. Diniz Marinho, The type I half-logistic family of distributions, J. Stat. Comput. Simul. 86 (2016), 707–728. 10.1080/00949655.2015.1031233Search in Google Scholar
[12] G. M. Cordeiro, M. Alizadeh, E. M. Ortega and L. H. V. Serrano, The Zografos–Balakrishnan odd log-logistic family of distributions: Properties and ypplications, Hacet. J. Math. Stat. 45 (2016), 1781–1803. 10.15672/HJMS.20159714145Search in Google Scholar
[13] G. M. Cordeiro, M. Alizadeh, G. Ozel, B. Hosseini, E. M. M. Ortega and E. Altun, The generalized odd log-logistic family of distributions: Properties, regression models and applications, J. Stat. Comput. Simul. (2016), 10.1080/00949655.2016.1238088. 10.1080/00949655.2016.1238088Search in Google Scholar
[14] G. M. Cordeiro, M. Alizadeh, M. H. Tahir, M. Mansoor, M. Bourguignon and G. G. Hamedani, The beta odd log-logistic generalized family of distributions, Hacet. J. Math. Stat. 45 (2016), 1175–1202. 10.15672/HJMS.20157311545Search in Google Scholar
[15] G. M. Cordeiro, E. M. Hashimoto and E. M. Ortega, McDonald Weibull model, Statistics 48 (2014), 256–278. 10.1080/02331888.2012.748769Search in Google Scholar
[16] I. Elbatal and G. Aryal, On the transmuted additive Weibull distribution, Austrian J. Stat. 42 (2013), 117–132. 10.17713/ajs.v42i2.160Search in Google Scholar
[17] R. C. Gupta, P. L. Gupta and R. D. Gupta, Modeling failure time data by Lehmann alternatives, Comm. Statist. Theory Methods 27 (1998), 887–904. 10.1080/03610929808832134Search in Google Scholar
[18] R. D. Gupta and D. Kundu, Generalized exponential distributions, Aust. N. Z. J. Stat. 41 (1999), 173–188. 10.1111/1467-842X.00072Search in Google Scholar
[19] M. N. Khan, The modified beta Weibull distribution, Hacet. J. Math. Stat. 44 (2015), 1553–1568. 10.15672/HJMS.2014408152Search in Google Scholar
[20] S. Kotz, C. D. Lai and M. Xie, On the effect of redundancy for systems with dependent components, IIE Trans. 35 (2003), 1103–1110. 10.1080/714044440Search in Google Scholar
[21] E. T. Lee and J. W. Wang, Statistical Methods for Survival Data Analysis, 3rd ed., Wiley, New York, 2003. 10.1002/0471458546Search in Google Scholar
[22] A. W. Marshall and I. Olkin, A new methods for adding a parameter to a family of distributions with application to the exponential and Weibull families, Biometrika 84 (1997), 641–652. 10.1093/biomet/84.3.641Search in Google Scholar
[23] F. Merovci, M. Alizadeh, H. M. Yousof and G. G. Hamedani, The exponentiated transmuted-G family of distributions: Theory and applications, Comm. Statist. Theory Methods (2016), 10.1080/03610926.2016.1248782. 10.1080/03610926.2016.1248782Search in Google Scholar
[24] F. Merovci and L. Puka, Transmuted Pareto distribution, ProbStat Forum 7 (2014), 1–11. Search in Google Scholar
[25] G. S. Mudholkar and D. K. Srivastava, Exponentiated Weibull family for analysing bathtub failure rate data, IEEE Trans. Reliab. 42 (1993), 299–302. 10.1109/24.229504Search in Google Scholar
[26] G. S. Mudholkar, D. K. Srivastava and M. Freimer, The exponentiated Weibull family: A reanalysis of the bus-motor-failure data, Technometrics 37 (1995), 436–445. 10.1080/00401706.1995.10484376Search in Google Scholar
[27] S. Nadarajah, The exponentiated Gumbel distribution with climate application, Environmetrics 17 (2005), 13–23. 10.1002/env.739Search in Google Scholar
[28] S. Nadarajah and A. K. Gupta, The exponentiated gamma distribution with application to drought data, Calcutta Statist. Assoc. Bull. 59 (2007), 29–54. 10.1177/0008068320070103Search in Google Scholar
[29] S. Nadarajah and S. Kotz, The exponentiated-type distributions, Acta Appl. Math. 92 (2006), 97–111. 10.1007/s10440-006-9055-0Search in Google Scholar
[30] Z. M. Nofal, A. Z. Afify, H. M. Yousof and G. M. Cordeiro, The generalized transmuted-G family of distributions, Comm. Statist. Simulation Comput. 46 (2017), 4119–4136. 10.1080/03610926.2015.1078478Search in Google Scholar
[31] M. W. A. Ramos, P. R. D. Marinho, G. Cordeiro, R. V. da Silva and G. G. Hamedani, The Kumaraswamy-G Poisson family of distributions, J. Stat. Theory Appl. 14 (2015), 222–239. 10.2991/jsta.2015.14.3.1Search in Google Scholar
[32] D. T. Shirke and C. S. Kakade, On exponentiated lognormal distribution, Int. J. Agric. Statist. Sci. 2 (2006), 319–326. Search in Google Scholar
[33] F. G. Silva, A. Percontini, E. de Brito, M. W. Ramos, R. Venancio and G. M. Cordeiro, The odd Lindley-G family of distributions, Austrian J. Stat. 46 (2017), 65–87. 10.17713/ajs.v46i1.222Search in Google Scholar
[34] M. H. Tahir, G. M. Cordeiro, A. Alzaatreh, M. Mansoor and M. Zubair, The logistic-X family of distributions and its applications, Comm. Statist. Theory Methods 45 (2016), 7326–7349. 10.1080/03610926.2014.980516Search in Google Scholar
[35] M. H. Tahir, M. Zubair, M. Mansoor, G. M. Cordeiro, M. Alizadeh and G. G. Hamedani, A new Weibull-G family of distributions, Hacet. J. Math. Stat. 45 (2016), 629–647. 10.15672/HJMS.2015579686Search in Google Scholar
[36] H. M. Yousof, A. Z. Afify, M. Alizadeh, N. S. Butt, G. G. Hamedani and M. M. Ali, The transmuted exponentiated generalized-G family of distributions, Pak. J. Stat. Oper. Res. 11 (2015), 441–464. 10.18187/pjsor.v11i4.1164Search in Google Scholar
[37] H. M. Yousof, A. Z. Afify, G. G. Hamedani and G. Aryal, The Burr X generator of distributions for lifetime data, J. Stat. Theory Appl., to appear. 10.2991/jsta.2017.16.3.2Search in Google Scholar
[38] H. M. Yousof, M. Rasekhi, A. Z. Afify, I. Ghosh, M. Alizadeh and G. G. Hamedani, The beta Weibull-G family of distributions: Theory, characterizations and applications, Pakistan J. Statist., to appear. Search in Google Scholar
© 2017 Walter de Gruyter GmbH, Berlin/Boston