a b s t r a c t
A new continuous distribution is introduced by compounding exponentiated exponential
and binomial distributions, named as exponentiated exponential binomial (EEB) distribution.
This distribution has the ability to model lifetime data with increasing, decreasing
and upside-down bathtub shaped failure rates. Moreover, the zero-truncated binomial distribution
used in compounding is overdispersed. Some properties of the distribution are
investigated. Estimation and inference procedure for the distribution parameters are discussed.
An application to real data demonstrates that the EEB distribution can provide a
better fit than a recent class of lifetime distributions.
1. Introduction
By various techniques, additional parameters can be introduced to expand families of distributions for added flexibility
to their hazard rates or to construct new models. Introducing a scale parameter leads to the accelerated life model, and
raising the cumulative distribution function (or survival function) to a fixed power introduces a parameter that leads to
the proportional hazards model. For instance, the family of Weibull distributions contains the exponential distribution and
is constructed by taking powers of exponential random variables. The gamma distributions also contain the exponential
distributions, and are constructed by taking powers of the Laplace transform. It is known that the distribution of a
sum of a fixed number of independent exponentially distributed random variables is a gamma and the distribution of a
minimum number of these variables is again an exponential distribution. Gupta and Kundu (1999) studied an exponentiated
exponential model that was constructed by the distribution of a maximum of independent exponential random variables
with a fixed sample size. Some authors used the minimum and the maximum of independent exponential random variables
to derive new families when the sample size is random. Marshall and Olkin (1997) considered a geometric distribution for the
sample size and derived some new extensions based on random minimum and maximum. Recently, Karlis (2009) and Kus
(2007) considered the exponential-Poisson distribution. For other related models see, e.g., Adamidis et al. (2005), Adamidis
and Loukas (1998) and Mudholkar and Srivastava (1993). In this paper, we introduce a new continuous distribution by
compounding exponentiated exponential and binomial distributions. This distribution offers a more flexible distribution for