Asymptotic Normality of the Estimators of the Parameters in Perturbed Gamma Model Using EM Algorithm.

Abstract

In this article it has been considered that the model which is the sum of two independent processes are gamma and uniform distribution and the mixture model is known as perturbed gamma model. Based on independent sample of the perturbed gamma model observed at irregular cases and when the missing data arises the expectation and maximization method is applied to estimate the unknown parameters of the model using the maximum likelihood estimator. A simulation study has been carried out to demonstrate the asymptotic normality of the estimators over three-dimensional graph in bivariate normal distribution using R software (maxLik/optim function).