التقدير الحصين لمعلمة القياس على المدى الجزئي

Robust estimation is considered an impotent subjects in statistic which is gives closed estimates from the classical estimates when the data contain outliers, as well as when the data are free of them. The classical standard deviation (SD) on of the most popular scale estimator which is influence by outliers, even by a single one. Therefore, in this paper we suggest (MADNa ) depending on the location estimator (MADa ) when (0 < α < 0.5). Then we compare between (MADNa ) and classical estimator (SD), when the data has normal distribution and contaminated normal distribution of one side with contamination ratio (e = 0.20), also we used Simulation techniques for generated data in different size and compare between (MADNa ) for (a = 0.1,0.25,0.4) with (SD) depending by the mean square error (MSE). It is shown through studding the simulation, that is used in this research, the (MADN0.4 ) is the best robust estimator of scale parameter (SD) for different sample size.