BAYESIAN PREDICTION BOUNDS FOR THE BURR TYPE-X MODEL

Based on a failure censored sample, the Bayesian prediction interval is obtained for the order statistic in the two sample and one sample cases when samples are assumed to be from the one-parameter Burr type X with p.d.f. f(x/theta) = 2x-theta-e(-x)2 (1-e(-x)2)theta-1 x greater-than-or-equal-to 0, theta > 0 where theta is assumed to be a realization of a random variable Kelvin. The conjugate and non-informative priors were assumed for Kelvin. As special cases, the Bayesian prediction interval is obtained for certain order statistics in the two cases above with some examples. A simulation study is also included.