Prediction Intervals for a Single Future Value of Normal and Non-Normal Variables

Authors

  • Sa-aat Niwitpong Department of Applied Statistics, Faculty of Applied Science, King Mongkut’s Institute of Technology North Bangkok,10800, Thailand.
  • Penpak Pinkoompee Department of Applied Statistics, Faculty of Applied Science, King Mongkut’s Institute of Technology North Bangkok,10800, Thailand.

Keywords:

bootstrap, coverage probability, expected length

Abstract

This paper investigates one-step-ahead prediction intervals for normal and non-normal variables. We propose studentized bootstrap and bootstrap percentile-t methods to construct one-step-ahead prediction intervals for normal and non-normal variables compared to a standard method. A minimum coverage probability 1- α and their expected lengths are used to select a preferable prediction interval. Monte Carlo simulation results show that not only does the bootstrap percentile-t give a better coverage probability but also give a shorter expected length.

Downloads

How to Cite

Niwitpong, S.- aat, & Pinkoompee, P. (2015). Prediction Intervals for a Single Future Value of Normal and Non-Normal Variables. Thailand Statistician, 4, 85–92. Retrieved from https://ph02.tci-thaijo.org/index.php/thaistat/article/view/34361

Issue

Section

Articles