Journal of Applied Sciences1812-56541812-5662Asian Network for Scientific Information10.3923/jas.2012.1313.1317Mohammed AhmedAl OmariIbrahimNoor AkmaAdamMohd BakriArasanJayanthi1220121212We consider the Weibull distribution which has been extensively
used in life testing and reliability studies of the strength of materials. The
maximum likelihood method is the usual frequentist approach in the parameter
estimate for parametric survival data. In this study, we divert from this platform
and use the Bayesian paradigm instead. The Jeffreys and extension of Jeffreys
prior with the squared loss function are considered in the estimation. The Bayes
estimates of the survival function and hazard rate of the Weibull distribution
with censored data obtained using Lindley’s approximation are then compared
to its maximum likelihood counterparts. The comparison criteria is the Mean
Square Error (MSE) and the performance of these three estimates are assessed
using simulations considering various sample sizes, several specific values
of Weibull parameters and several values of extension of Jeffreys prior. The
maximum likelihood estimates of survival function and hazard rate are more efficient
than their Bayesian counterparts, however, the extension of Jeffreys is better
than the maximum likelihood for certain conditions.]]>Assoudou, S. and B. Essebbar,2003Ahmed, A.O.M., H.S. Al-Kutubi and N.A. Ibrahim,2010Ahmed, A.O.M. and N.A. Ibrahim,2011Cohen, A.C. and B. Whitten,1982Hahn, J.,2004Hossain, A.M. and W.J. Zimmer,2003Klein, J.P. and M.L. Moeschberger,2003Preda, V., A. Constantinescu and E. Panaitescu,2010Singh, R., S.K. Singh, U. Singh and G.P. Singh,2008Singh, U., P.K. Gupta and S. Upadhyay,2002Singh, U., P.K. Gupta and S.K. Upadhyay,2005Sinha, S.K.,1986Sinha, S. K. and J.A. Sloan,1988Soliman, A.A., A.H. Abd Ellah and K.S. Sultan,2006