On the Independence of Jeffreys’ Prior For Truncated-Exponential Skew-Symmetric Models

We study the jeffreys’ prior of the skewness parameter of truncated-exponential skew-symmetric distributions (tessds). we show that this prior is symmetric, improper, and with tails o(|λ| ^−1 ). while jeffreys’ prior is improper, as we have shown, it yields a proper posterior distribution for some densities. we also calculate the independent jeffreys’ prior for the case of unknown location and scale parameters and show that the corresponding posterior distribution is proper. a simulation study using monte carlo methods is presented to compare the efficiency of bayesian estimators in tessd with azzalinis’ skew models by computing the bias and the mean square error under square error loss and linex loss functions. the results show the superiority of the bayesian estimators in tessd versus bayesian estimators in azzalinis’ skew models.