Approximations for the inverse gaussian probabilities and percentiles

Abstract

Recent years have seen accelerated use of the inverse Gaussian distribution to model asymmetric data in many areas of applied research. Yet, the inverse Gaussian percentiles and probabilities needed in analyses are not readily available, and a variety of two moment approximations have been employed for the purpose. In this note we propose some new approximations and provide guidelines for their use.

Keywords:

• skewuess reduction
• lognormal
• extreme tails