Asymptotic Sampling Distribution of Inverse Coefficient of Variation and its Applications: Revisited
No Thumbnail Available
Authors
Albatineh, Ahmad
Kibria, Golam
Zogheib, Bashar
Issue Date
2014
Type
Journal Article
Peer-Reviewed
Peer-Reviewed
Language
Keywords
Alternative Title
Abstract
Sharma and Krishna (1994) derived mathematically an appealing asymptotic confidence interval for the population signal-to-noise ratio (SNR). In this paper, an evaluation of the performance of this interval using monte carlo simulations using randomly generated data from normal, log-normal, $\chi^2$, Gamma, and Weibull distributions three of which are discussed in Sharma and Krishna (1994). Simulations revealed that its performance, as measured by coverage probability, is totally dependent on the amount of noise introduced. A proposal for using ranked set sampling (RSS) instead of simple random sampling (SRS) improved its performance. It is recommended against using this confidence interval for data from a log-normal distribution. Moreover, this interval performs poorly in all other distributions unless the SNR is around one.
Description
Citation
Publisher
License
Journal
Volume
2
Issue
1