الفهرس 
New biased estimators for seemingly unrelated regression equations model in case of incomplete data
Table of ContentsOnly 14 pages are availabe for public view
 |  | from | 130 |  |  |
| | | from | 130 | | |
Abstract
SURE model which was first proposed by Zellnerin 1962 became one of the most important regression models in the application aspects. In this thesis, we will estimate SURE model which suffers from missing values multicollinearity issues. This study is different in many aspects from other studies which estimated SURE model that handling the missing values and the multicollinearity under different factors such as sample size, percent of missing values, number of equations, number of explanatory variables and degree of correlation between the explanatory variables. All the previous factors are studied on different levels to know the effect of each level on the efficiency of the proposed estimators. For handling missing values in SURE model, four imputation approaches were introduced which are regression imputation; PMM; EM algorithm and MCMC. After that, we will handle multicollinearity problem by using Liu-Type estimation with Two-parameters and ridge estimation to estimate the bias. The Monte Carlo simulation study was done with 1000 replications.The results revealed that the EM-Liu type estimators were the best for imputing the missing values and handling the multicollinearity over all the factor and all levels in the terms of TMSE