ASYMPTOTIC EFFICIENCY OF ESTIMATORS UNDER ASYMMETRIC LOSS FUNCTIONS

抄録

The efficiency of estimators is discussed based on risk function for asymmetric loss functions. Specifically, an asymptotically lower risk bound is established for appropriate classes of estimators. The asymmetry brings about the shift of location of the loss function. Furthermore, under certain assumptions, the optimal estimators that attain the bound are constructed by utilizing the estimating function with asymptotic bias.