Double K-class estimators under multivariate t-errors /
مع أخطاء لها توزيع ت متعدد المتغيرات (k) مجموعة ثنائى
Abdelraheim Mohamed Abdelraheim Mohamed ; Supervised Amany Mossa Mohamed , Elhoussainy Abdelbar Rady
- Cairo : Abdelraheim Mohamed Abdelraheim Mohamed , 2017
- 105 Leaves ; 30cm
Thesis (Ph.D.) - Cairo University - Institute of Statistical Studies and Research - Department of Statistics and Econometrics
The main idea of the thesis is based on analysis of exact and asymptotic bias, moment matrix and risk function for double k-class estimators in general linear regression model under multivariate normal errors. An analysis of exact and asymptotic bias, moment matrix and risk function for double k- class estimators with multivariate t-errors under quadratic loss function. It contribute a better understanding of the performance of double k-class estimators and the comparison between double k-class estimators and ordinary least squares estimators for the same linear regression model. An analysis of exact and asymptotic risk function for double k-class estimators with multivariate t-errors under general quadratic loss function, demonstrate that the double k-class estimators dominates the ordinary least squares and the characterizing scalars which nearly minimize the risk of the estimator, and examine the behavior of risk function under Multivariate t-errors. The thesis used a simulation study to compare the performance of the family of double k-class estimators, and introduces an empirical study for the application on the Egyptian Stock Market to compare the performance of the family of double k-class estimators. It worth mention, that the results have been published in three papers. The first at. The 50th, annual conference on statistics and computer science and operation research, ISSR-Cairo university, Egypt in 2015 under title some properties of double k-class estimators in linear regression under multivariate t-errors
Bias estimators Double k-class estimators Risk function