An alternative estimator for the shape parameter in the negative binomial distribution

摘要

Best and Gipps (1974) showed that the negative binomial distribution can be approximated closely by a gamma distribution. Using the moment estimators for the parameters in the gamma distribution, we use the approach of Best and Gipps to propose an alternative estimator for the shape parameter a which, while biased, appears to perform better in many situations, both in terms of mean square error and percentile measure, than its key competitor, namely the method of moments estimator.