الفهرس 
Estimation of linear regression models with incomplete data
Table of ContentsOnly 14 pages are availabe for public view
 |  | from | 105 |  |  |
| | | from | 105 | | |
Abstract
In linear regression, the incomplete values occur in sample. Many different methods to estimate parameters in the regression model are: (1)- The Ordinary Least Squares (OLS) (2)-Maximum Likelihood (ML) (3)- Weighted Least Squares (WLS). However, the validity of these approaches are decreased in making inference because the size of the sample reduces. In this thesis, when missing values occur in the independent variable, we perform a simulation study of imputation based on procedures and indicate that missing value should be filled by the mean and regression imputation methods, which have been more efficient than other methods. Different methods of estimating the missing values have been used. Also, different methods of estimating the parameters of the regression model have been used. In case of having different sample sizes , different variances of errors and different proportions of missing values when the missing in the independent variables, the Maximum Likelihood (ML) method is a better one respect to the Original Least Squares (OLS) and Weighted Least Squares (WLS) methods with mean method (as method of estimating the missing values). But when the missing in both the independent and dependent variables, the Maximum Likelihood (ML) method is also, a better one respect to the Original Least Squares (OLS) and Weighted Least Squares (WLS) methods with mean method (as method of estimating the missing values)