Natchaya Ratasukharom.. Inferences for the parameters of Birnbaum-Saunders distributions. ปริญญาเอก(Applied Statistics). King Mongkut's University of Technology North Bangkok. Central Library. : King Mongkut's University of Technology North Bangkok, 2024.
Inferences for the parameters of Birnbaum-Saunders distributions
Abstract:
The Birnbaum-Saunders (BirSau) distribution, which is asymmetric, can be
transformed into a symmetric normal distribution. Thus, the BirSau distribution is
useful for modeling data with values greater than zero or positive observation. When
data includes both positive and zero observations, the zero-inflated Birnbaum-Saunders
(ZI-BirSau) distribution is used. This thesis uses confidence intervals for single-
parameter and two-parameter simulations of the BirSau and ZI-BirSau distributions.
The thesis aims to construct confidence intervals for the mean of the BirSau
distributions and the mean and variance of the ZI-BirSau distributions. Simultaneous
confidence intervals for the difference of means of the ZI-BirSau distributions are also created. The best-performing confidence interval is selected based on its coverage probability (CP), which should be greater than or close to the nominal confidence level of 0.95, and the shortest average width (AW) as determined by Monte Carlo simulations. Finally, wind speed data are given to illustrate the confidence intervals.