Efficient estimation of negative binomial parameters using empirical La Place transform

Abstract

A new method, based on empirical La Place transform, was developed to find asymptotically efficient estimates of negative binomial distribution parameters. These estimates were found fairly close to those found by the method of maximum likelihood. Efficiencies over 95 percent were obtained. The method was tested with a set of data, generated by computer, and found to be satisfactory except in a few cases. Maximum likelihood also fails to be satisfactory in these cases.