Estimation of Parameters for a Generalized Hypergeometric Distribution Generated by Birth-Death Process

Abstract

In this paper, we consider a three parametric regularly varying generalized Hypergeometric distribution which have been generated by Birth-Death process for describing phenomena in bioinformatics (Danielian and Astola, 2006). Under satisfying some conditions, we obtain the system of likelihood equations which its solution coincides with the maximum likelihood estimators. The given maximum likelihood estimators are the same as some moment estimators. Moreover, an approximate computation of the maximum likelihood estimations for the unknown parameters is given. Using MCMC, simulation studies are proposed. Finally, in order to present applications, some real data sets in bioinformatics are fitted with the model. Based on some important criterions, this model is compared with four other discrete distributions in bioinformatics. We see that the generalized Hypergeometric distribution provides a better fit than four other discrete distributions../files/site1/files/61/9(1).pdf

Keywords

• generalized hypergeometric distribution
• birth-death process
• bioinformatics
• maximum likelihood estimation (mle)
• markov chain monte carlo (mcmc).