摘要:Recently Kundu and Gupta (2009) introduced a bivariate generalized exponential dis- tribution, whose marginals are generalized exponential distributions. The bivariate gen- eralized exponential distribution is a singular distribution, similarly as the well known bivariate Weibull distribution. The corresponding two singular bivariate distributions functions have very similar joint probability density functions. In this paper, we consider the discrimination between these two bivariate distribution functions. The di.erence of the maximized log-likelihood functions is used in discriminating between the two dis- tribution functions. The asymptotic distribution of the test statistic has been obtained and it can be used to compute the asymptotic probability of correct selection. Monte Carlo simulations are performed to study the e.ectiveness of the proposed method. One data set has been analyzed for illustrative purposes.
关键词:Asymptotic distribution ¢ EM algorithm ¢ Likelihood ratio test;¢ Maximum likelihood ¢ Monte Carlo simulations ¢ Probability of correct selections.