On the Prior Selection for Variance of Normal Distribution

by Muhammad Aslam and Abdul Haq.

Abstract: This study provides a comparison of informative priors for unknown parameter variance of the normal distribution with known parameter mean. We have assumed four possible informative priors: Inverse chi-squared, Inverted Gamma, Levy and Gumbel type-II distributions for the variance of the normal distribution. The objective of study is to select suitable prior for normal variance. The comparison is based upon the posterior variances, coefficients of skewness, also on the basis of posterior predictive variances. Finally a suitable prior for the unknown variance of a normal distribution is recommended.

Key Words: Density kernel, Informative prior, Posterior distribution, Posterior predictive distribution, Prior distribution

Muhammad Aslam, aslamsdqu@yahoo.com
Abdul Haq, aaabdulhaq@yahoo.com

Editor: Bhattacharjee, Dibyojyoti, lecturer_dibyo@yahoo.com lecturer_dibyo@yahoo.com

READING THE ARTICLE: You can read the article in portable document (.pdf) format (234776 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 2295 times since MAY 6, 2009.

Return to the InterStat Home Page.