Abstract
Multivariate compound Poisson distribution is obtained by m p-independent Poisson distributions with a mixing distribution. marginal, conditional probability mass function and moments of comp Poisson distribution when the mixing parameter has a gamma distribu are obtained. The parameters of the mixed distribution are then estirr under a normalized squared error loss. The improved estimators obtained over the unbiased estimators for ' p> = 2, by solving ce difference inequality. Also robustness of the Clevenson-Zidek type estin is shown when the p-distributions are dependent through the m process. Risk simulations studies are obtained to show the amoui improvements.