Conditional information for an inverse Gaussian distribution with known coefficient of variation |
| |
Authors: | Katuomi Hirano Kōsei Iwase |
| |
Affiliation: | (1) The Institute of Statistical Mathematics, 4-6-7 Minami-Azabu, Minato-ku, 106 Tokyo, Japan;(2) Department of Applied Mathematics, Faculty of Engineering, Hiroshima University, 724 Higashi-Hiroshima, Japan |
| |
Abstract: | Conditional inference about a mean of an inverse Gaussian distribution with known coefficient of variation is discussed. For a random sample from the distribution, sufficient statistics with respect to the mean parameter include an ancillary statistic. The effects of conditioning on the ancillary statistic are investigated. It is shown that the model provides a good illustration of R. A. Fisher's recommendation concerning use of the observed second derivative of the log likelihood function in normal approximations.This work was started while Ksei Iwase was visiting the Institute of Statistical Mathematics in Spring, 1987, and was partly supported by the ISM Cooperative Research Program (88-ISM·CRP-7), and by Scientific Research Fund No. 62540173 from the Ministry of Education, Science and Culture of Japan. |
| |
Keywords: | Ancillary statistic coefficient of variation conditional inference inverse Gaussian distribution sample information |
本文献已被 SpringerLink 等数据库收录! |
|