Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
416744 | Computational Statistics & Data Analysis | 2006 | 12 Pages |
Abstract
In this paper, we propose two different confidence regions for a mean vector under general conditions when the population of interest is nonnormal and the sample size is moderate. It is shown that both confidence regions will improve the accuracy of the approximation in the sense that the coverage error is of order o(n-1)on-1. Monte Carlo examples are demonstrated to show the performance with comparison to the classical methods.
Related Topics
Physical Sciences and Engineering
Computer Science
Computational Theory and Mathematics
Authors
Jin Xu, Arjun K. Gupta,