THE WEIBULL GENERALIZED FLEXIBLE WEIBULL EXTENSION DISTRIBUTION

Abdelfattah Mustafa, Beih S. El-Desouky, Shamsan AL-Garash Department of Mathematics, Faculty of Science, Mansoura University, Mansoura 35516, Egypt Abstract: This paper introduces a new four parameters model called the Weibull Generalized Flexible Weibull extension (WGFWE) distribution which exhibits bathtub-shaped hazard rate. Some of it’s statistical properties are obtained including ordinary and incomplete moments, quantile and generating functions, reliability and order statistics. The method of maximum likelihood is used for estimating the model parameters and the observed Fisher’s information matrix is derived. We illustrate the usefulness of the proposed model by applications to real data.


Introduction
The Weibull distribution is a highly known distribution due to its utility in modelling lifetime data where the hazard rate function is monotone Weibull (1951).In recent years new classes of distributions were proposed based on modifications of the Weibull distribution to cope with bathtub hazard failure rate Xie and Lai (1995).Exponentiated modified Weibull extension while the cumulative distribution function (cdf) is given by The survival function is given by the equation and the hazard function is Weibull distribution introduced by Weibull (1951), is a popular distribution for modeling phenomenon with monotonic failure rates.But this distribution does not provide a good fit to data sets with bathtub shaped or upside-down bathtub shaped (unimodal) failure rates, often encountered in reliability, engineering and biological studies.Hence a number of new distributions modeling the data in a better way have been constructed in literature as ramifications of Weibull distribution.Marcelo  .Consider that the variability of this odds of death is represented by the random variable X and assume that it follows the Weibull model with scale a and shape b.We can write Which is given by Eq. ( 6).works as a corrected factor for the hazard rate function of the baseline model ( 6) can deal with general situation in modeling survival data with various shapes of the hazard rate function.By using the power series for the exponential function, we obtain substituting from Eq.( 10) into Eq.( 7), we get Using the generalized binomial theorem we have Inserting Eq. ( 12) in Eq. ( 11), the Weibull-G family density function is This paper is organized as follows, we define the cumulative, density and hazard functions of the Weibull-G Flexible Weibull Extension (WGFWE) distribution in Section 2. In Sections 3 and 4, we introduced the statistical properties include, quantile function skewness and kurtosis, rth moments and moment generating function.The distribution of the order statistics is expressed in Section 5.The maximum likelihood estimation of the parameters is determined in Section 6.Real data sets are analyzed in Section 7 and the results are compared with existing distributions.Finally, Section 8 concludes.

The Weibull-G Flexible Weibull Extension Distribution
In this section we studied the four parameters Weibull-G Flexible Weibull Extension (WGFWE) distribution.Using G(x) and g(x) in Eq. ( 13) to be the cdf and pdf of Eq. ( 6) and Eq. ( 7).The cumulative distribution function cdf of the Weibull-G Flexible Weibull Extension distribution (WGFWE) is given by The pdf corresponding to Eq. ( 14) is given by where x > 0 and , α, β > 0 are two additional shape parameters.From Figures 1-5, the WGFWE distribution is unimodal distribution, the hazard rate is decreasing, increasing and constant, decreasing reversed hazard rate and survival function.

Statistical Properties
In this section, we study the statistical properties for the WGFWE distribution, specially quantile function and simulation median, skewness, kurtosis and moments.

Quantile and simulation
The quantile   of the WGFWE(a, b, α, β) random variable is given by.
Using the distribution function of WGFWE, from ( 14), we have Where So, the simulation of the WGFWE random variable is straightforward.Let U be a uniform variate on the unit interval (0, 1).Thus, by means of the inverse transformation method, we consider the random variable X given by Since the median is 50% quantile, we obtain the median M of X by setting q = 0.5 in Eq. ( 21).

The Skewness and Kurtosis
The analysis of the variability Skewness and Kurtosis on the shape parameters α, β can be investigated based on quantile measures.The short comings of the classical Kurtosis measure are well-known.The Bowely's skewness based on quartiles is given by, Kenney and Keeping (1962).and the Moors Kurtosis based on quantiles, Moors (1998) where q(.) represents quantile function.

The Moment
In this subsection we discuss the rth moment for WGFWE distribution.Moments are important in any statistical analysis, especially in applications.It can.be used to study the most important features and characteristics of a distribution (e.g.tendency, dispersion, skewness and kurtosis).
Theorem 1.If X has WGFWE (a, b, α, β) distribution, then the rth moments of random variable X, is given by the following Proof.We start with the well known distribution of the rth moment of the random variable X with probability density function f (x) given by Substituting from Eq. ( 1) and Eq. ( 2) into Eq.( 13

Order Statistics
In this section, we derive closed form expressions for the probability density function of the rth order statistic of the WGFWE distribution.Let

Asymptotic confidence bounds
In this section, we derive the asymptotic confidence intervals when a, b, α > 0 and β > 0 as the MLEs of the unknown parameters a, b, α > 0 and β > 0 can not be obtained in closed forms, by using variance covariance matrix  −1 see Lawless (2003), where  −1 is the inverse of the observed information matrix which defined as follows The second partial derivatives included in I are given as follows.
Table 2 gives MLEs of parameters of the WGFWE distribution and K-S Statistics.The values of the log-likelihood functions, AIC, AICC, BIC and HQIC are in Table 3.
We find that the WGFWE distribution with four parameters provides a better fit than the previous models flexible Weibull (FW), Weibull (W), linear failure rate (LFR), exponentiated Weibull (EW), generalized linear failure rate (GLFR) and exponentiated flexible Weibull (EFW).It has the largest likelihood, and the smallest K-S, AIC, AICC, BIC and HQIC values among those considered in this paper.
The nonparametric estimate of the survival function using the Kaplan-Meier method and its fitted parametric estimations when the distribution is assumed to be WGFWE, FW , W , LFR, EW , GLFR and EFW are computed and plotted in Figure 6.To show that the likelihood equation have unique solution, we plot the profiles of the loglikelihood function of a, b, α and β in Figures 7 and 8.The nonparametric estimate of the survival function using the Kaplan-Meier method and its fitted parametric estimations when the distribution is assumed to be WGFWE, FW, W, MW, RAW and EW are computed and plotted in Figure 9.

Summary
A new distribution, based on Weibull-G Family distributions, has been pro-posed and its properties are studied.The idea is to add parameter to a flexible Weibull extension distribution, so that the hazard function is either increasing or more importantly, bathtub shaped.Using Weibull generator component, the distribution has flexibility to model the second peak in a distribution.We have shown that the Weibull-G flexible Weibull extension distribution fits certain well-known data sets better than existing modifications of the Weibull-G family of probability distribution.
et al. (2014) introduced and studied generality a family of univariate distributions with two additional parameters, similarly as the extended Weibull, Gurvich et al. (1998) and Gamma-families, Zografos and Balakrshnan (2009), using the Weibull generator applied to the odds ratio () 1−() .If G(x) is the baseline cumulative distribution function (cdf) of a random variable, with probability density function (pdf) g(x) and the Weibull cumulative distribution function is with parameters a and b are positive.Based on this density, by replacing x with ratio () 1−() .The cdf of Weibull-generalized distribution, say Weibull-G distribution with two extra parameters a and b, is defined by Marcelo et al. (2014) where G(x; θ) is a baseline cdf, which depends on a parameters vector θ.The corresponding family pdf becomes A random variable X with pdf (7) is denoted by X distributed Weibll-G(a, b, θ), x ∈ R, a, b > 0. The additional parameters induced by the Weibull generator are sought as a manner to furnish a more flexible distribution.If b = 1, it corresponds to the exponential-generator.An interpretation of the Weibull-G family of distributions can by given as follows (Corollary, Cooray (2006)) is a similar context.Let Y be a lifetime random variable having a certain continuous G distribution.The odds ratio that an individual (or component) following the lifetime Y will die (failure) at time x is () 1−() The survival function of the Weibull-G family is given by and hazard rate function of the Weibull-G family is given by where h(x; θ) = (;) 1−(;) .The multiplying quantity •(;)[(;)] −1 [1−(;)] The survival function S(x), hazard rate function h(x), reversed-hazard rate function r(x) and cumulative hazard rate function H(x) of X ∼ WGFWE(a, b, α, β)are given by respectively, x > 0 and a, b, α, β > 0. Figures (1-5) display the cdf, pdf, survival, hazard rate and reversed hazard rate function of the WGFWE(a, b, α, β) distribution for some parameter values.

Figure 1 :
Figure 1:The cdf for different values of parameters.

Figure 2 :
Figure 2:The pdf for different values of parameters
1: ,  2: …,  : denote the order statistics obtained from a random sample  1 ,  2 , • • • ,   which taken from a continuous population with cumulative distribution function cdf F (x;) and probability density function pdf f (x; φ),then the probability density function of  : is given by where f (x; φ), F (x;)are the pdf and cdf of WGFWE(φ) distribution given by Eq. (15) and Eq.(14) respectively, φ = (a, b, α, β) and B(., .) is the Beta function, also we define first order The maximum likelihood estimation of the parameters are obtained by differentiating the loglikelihood function L with respect to the parameters a, b, α and β and setting the result to zero, we have the following normal equations.where D  = exp{α  − }.The MLEs can be obtained by solving the nonlinear equations previous, (35)-(38), numerically for a, b, α and β.

Figure 6 :
Figure 6: The Kaplan-Meier estimate of the survival function for the data

Figure 7 :
Figure 7: The profile of the log-likelihood function of a, b.

Figure 8 :
Figure 8: The profile of the log-likelihood function of α, β.

Figure 9 :
Figure 9: The Kaplan-Meier estimate of the survival function for the data.

Figures 10 and 11
Figures 10 and 11 give the form of the hazard rate and CDF for the WGFWE, FW, W, MW, RAW and EW which are used to fit the data after replacing the unknown parameters included in each distribution by their MLE.

Figure 10 :
Figure 10: The Fitted hazard rate function for the data.

Figure 11 :
Figure 11:The Fitted cumulative distribution function for the data.