The Birnbaum-Saunders (BS) model is a two-parameter, unimodal distribution with positive skewness and non-negative support. This model has attracted considerable attention in the statistical literature during the past decade. The interest for the BS model is due to its physical theoretical arguments, its attractive properties and its relationship with the normal model. However, no much attention has been payed in parameterizations of the BS distribution distinct to that proposed originally. In this paper, we study different parameterizations of the BS distribution. Specifically, we analyze the properties of these parameterizations, use the maximum likelihood method for estimating the corresponding parameters and carry out a simulation study for detecting their performance.
|Number of pages||26|
|Journal||Pakistan Journal of Statistics|
|State||Published - 1 Jan 2012|
- Confidence intervals
- Coverage probabilities
- Delta ML methods
- Monte Carlo simulations