الفهرس Only 14 pages are availabe for public view
 |  | from | 145 |  |  |
| | | from | 145 | | |
Abstract
The major purpose for this thesis is to study the estimation of variance components when part of information is completely missing using modified minimum variance quadratic unbiased estimation (MMIVQUE) method. The estimators of variance components are derived in unbalanced (balanced) one-way random model and two-way nested random model for data with complete information and data with completely missing information. The study included five chapters as follows: Chapter one included the concept of variance components, variance components models and methods of variance components estimation. Chapter two presents difference between data with missing values and data with missing information. Literature review of estimation of variance components for data with missing values and missing information is introduced. In chapter three, the estimators of variance components are derived in one-way random model for unbalanced (balanced) data with complete information and data with completely missing information by modified MIVQUE. In chapter four, the estimators of variance components are derived in two-way nested random model for unbalanced (balanced) data with complete information and completely missing information by modified MIVQUE. Finally, chapter five consists of two sections: Simulation study and application study. The variance components are estimated in different models and to compare the estimates by mean squared error, absolute Bias, probability of getting negative estimates and relative efficiency