摘要

In this article, a new distribution, called a beta power transformed Generalized Exponential Distribution (aPTGE) is proposed. The distribution has simple closed forms pdf and cdf. Several properties of the aPTGE distribution are studied including unimodality, moments, conditional moments, stochastic ordering and Shannon entropy. The distribution is capable of monotonically increasing, decreasing, bathtub and upside down shaped hazard rates. Furthermore, we discussed several methods of estimation for estimating the aPTGE(alpha, beta, lambda) parameters. In particular, we compare, through simulation study, the performance of the maximum likelihood estimation, maximum product spacing estimation, ordinary and weighted least-squares estimation methods. Also, the average interval lengths and coverage rates of 95% confidence intervals, including bootp and boot-t confidence intervals for different parameter values are provided. A real data set is used to illustrate the applicability of the proposed distribution.