AISM 54, 840-847
© 2002 ISM
(Received April 5, 2001; revised July 16, 2001)
Abstract. In this paper, the more convenient estimators of both parameters of the gamma distribution are proposed by using its characterization, and shown to be more efficient than the maximum likelihood estimator and the moment estimator for small samples. Furthermore, the distribution of the square of the sample coefficient of variation is obtained by computer simulation for some various values of the parameters and sample size, and thus the simulated confidence interval of its shape parameter is established.
Key words and phrases: Sample coefficient of variation, shape parameter, moment estimator, gamma distribution.
Source (TeX , DVI )