Advances and Applications in Statistics
Volume 33, Issue 1, Pages 1 - 21
(March 2013)
|
|
EFFICIENT METHODS ON CONFIDENCE INTERVALS FOR PREDICTION INTERVALS
Mei Ling Huang, Wai Kong Yuen and Miao Zhang
|
Abstract: Prediction intervals have many applications in economics, business, health studies, engineering, science and social science. In this article, we study confidence intervals for prediction intervals of the future value of a random variable based on quantile estimations. There are theoretical difficulties in this problem with few methods in the literature. We propose three methods based on weight functions combined with bootstrapping. The results of Monte Carlo simulations confirm that the proposed methods improve the efficiencies in mean square error and probability coverage relative to the existing methods. We also study a real world example which shows the improved efficiencies of the proposed methods. |
Keywords and phrases: bootstrapping, efficiency, HD estimator, kernel density estimation, L‑statistics, probability coverage, quantile estimation, weighted empirical distribution function. |
|
Number of Downloads: 348 | Number of Views: 1085 |
|