A Generalized Family of Exponentiated Composite Distributions
In this paper, we propose a new class of distributions by exponentiating the random variables associated with the probability density functions of composite distributions. We also derive some mathematical properties of this new class of distributions including the moments and the limited moments. Specifically, two special models in this family are discussed. Two real data sets were chosen to assess the performance of these two special exponentiated composite models. When fitting to these two data sets, theses two special exponentiated composite distributions demonstrate significantly better performance compared to the original composite distributions.