مقارنة بين مقدرات S و أصغر موجه تباين معاد الأوزان (RMVV)‎ لتقدير معلمتي الموقع و القياس للتوزيع الطبيعي متعدد المتغيرات : بحث تطبيقي

Abstract EN

In this research, robust and efficient capabilities of the mean wave (M) and the covariance and variance matrix (∑) of the multivariate normal distribution were found when the hypotheses related to this distribution were penetrated and when the outliers were present in the sample under study, by using the method (smaller wave).

The variance of the weights equation and its comparison with the method of estimators of S, where the simulation method was used to compare the two methods and through the mean squares of error (MSE) of the averages wave and the variance matrix and the covariance ∑ to reach the best method of estimation, and the simulation results showed that the method of smaller wave variation of weights is the best because it achieved the lowest mean squares of error (MSE) than the other method, and with different sample sizes and contamination ratios, and on the application side, both methods were used on the data of kidney patients and the results showed an agreement with the experimental side, where the method achieved (the smallest wave contrast re-weighted) the preference in estimation through the MSE.