Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1149651 | Journal of Statistical Planning and Inference | 2012 | 16 Pages |
Abstract
We study the asymptotic behavior of the weighted sum of correlated chi-squared random variables. Both chi-squared and normal distributions are proved to approximate the exact distribution. These two approximations are established by matching the first two cumulants. Simulation comparison is made to study the performance of two approximations numerically. We find that the chi-squared approximation performs better than the normal one in the study.
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
Li-Ling Chuang, Yu-Shan Shih,