Research and Analysis on Optimizing Five Empirical DistributionFunctions in Mean once Distance

摘要

This article introduces five empirical distribution functions of Mean once Distance under the condition of absolute loss. To these empirical distribution functions, we applies knowledge of statistical decision and thoughts of risk function, and do simulation, comparisons and analysis, then obtain the most consistent optimal empirical distribution function W(Hn,F) in mean once distance. Thus, a new statistics Wn and testing sequence are established. Finally, an example simulated by software MATLAB is set to indicate that the empirical distribution function in mean once distance is available.