Parameter Estimation Algorithms for the Hyper-Gamma Distribution Class |
| |
Authors: | Helmut P Dudel Charles E Hall Jr Siegfried H Lehnigk |
| |
Institution: | U.S. Army Missile Command Research, Development, and Engineering Center , AMSMI-RD-RE-OP , 35898–5246, AL, Redstone Arsenal |
| |
Abstract: | Microcomputer-based algorithms for the estimation of the parameters shift, scale, initial and terminal shape of the hyper–Gamma distribution class are presented. They are based on the moment equations and on the logarithmic likelihood function (LLF) associated with the hyper-Gamma density. The maximum–likelihood approach is implemented by means of the derivative equations resulting from the LLF and, independently, by means of direct optimization of the LLF. Program options include estimation of (i) four parameters, (ii) three parameters (shift known), and (iii) two parameters (shift known, initial shape zero). A program diskette with user's guide will be made available upon request. |
| |
Keywords: | four three and two parameter densities moment estimates maximum-likelihood estimates by derivative equations and by optimization numerical algorithms |
|
|