A0800
Title: On parameter estimation for generalized inverse Gaussian distribution
Authors: Hideki Nagatsuka - Chuo University (Japan) [presenting]
Shunsuke Kaneko - Chuo University (Japan)
Abstract: The generalized inverse Gaussian (GIG) distribution, provided by Halphen in 1941 and then developed by Barndorff-Nielsen, is a generalized distribution of the inverse Gaussian and gamma distributions. The GIG distribution has some desired properties. For example, any GIG distribution with a non-positive power parameter is the first hitting time to level 0 for a time-homogeneous diffusion process, which implies the potential use of this distribution as a lifetime distribution. The GIG distribution has infinite divisibility, which suggests that a Levy process can be constructed based on the GIG distribution. Some challenging problems are addressed in parameter estimation for the GIG distribution, and some applications of this distribution are introduced.
