Slashed generalized exponential distribution

Abstract

In this paper, we introduce an extension of the generalized exponential distribution, making it more robust against possible influential observations. The new model is defined as the quotient between a generalized exponential random variable and a beta distributed random variable with one unknown parameter. The resulting distribution is a distribution with greater kurtosis than the generalized exponential distribution. Probability properties of the distribution such as moments and asymmetry and kurtosis are studied. Likewise, statistical properties are investigated using the method of moments and the maximum likelihood approach. Two real data analyses are reported illustrating better performance of the new model over the generalized exponential mode