Bayes Reliability Measures of Lognormal and Inverse Gaussian Distributions under ML-II e-contaminated Class of Prior Distributions

  • Pankaj Sinha University of Delhi, Delhi
  • J. Prabha University of Delhi, Delhi
DOI: https://doi.org/10.14429/dsj.60.486
Keywords: Bayes reliability, lognormal distribution, inverse Gaussian distribution, Bayesian methodology

Abstract

ML-II e-contaminated class of priors are employed to study the sensitivity of Bayes reliability measures for a lognormal (LN) distribution and inverse Gaussian (IG) distribution to mis-specification in the prior is employed. The numerical illustrations suggest that reliability measures of both the distributions are not sensitive to moderate amount of mis-specification in prior distributions belonging to the class of ML-II e-contaminated.

Defence Science Journal, 2010, 60(4), pp.442-450, DOI:http://dx.doi.org/10.14429/dsj.60.486

Author Biographies

Pankaj Sinha, University of Delhi, Delhi
Received his PhD in Bayesian Econometrics from University of Delhi. Presently he is an Assoc Prof at Faculty of Management Studies, University of Delhi.
J. Prabha, University of Delhi, Delhi
Obtained her MPhil in Statistics from University of Delhi. Presently she is pursuing PhD in Statistics on empirical Bayes approach to modelling financial volatility, asset pricing and portfolio selection from University of Delhi.
Published
2010-07-09
How to Cite
Sinha, P., & Prabha, J. (2010). Bayes Reliability Measures of Lognormal and Inverse Gaussian Distributions under ML-II e-contaminated Class of Prior Distributions. Defence Science Journal, 60(4), 442-450. https://doi.org/10.14429/dsj.60.486
Issue
Vol 60 No 4: Machine Learning and Pattern Recognition
Section
Research Papers
Copyright (c) 2016 Defence Science Journal

 Where otherwise noted, the Articles on this site are licensed under Creative Commons License: CC Attribution-Noncommercial-No Derivative Works 2.5 India