Lucknow University, Lucknow, India
Robustness plays an important role in Bayesian Inference where estimators are improved with help of sample and prior information. Huber (1969), Andrews (1974) and Bierens (1981) suggested robustness in different classical and Bayesian models. Malinvaud (1970) has stated that robust methods are those statistical procedures which show little sensitivity to the assumptions on the stochastic variables of the model for which they are conceived. In the present paper, we have studied ‘Robustness’ on Bayes estimates of Log-Normal distribution by considering a ‘quasi-prior’ for the parameters involved.
