الفهرس 
A Multiple Objectives Programming Approach for Multivariate Calibration Estimation in Stratified Random Sampling with Application
AbstractOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
Calibration estimation is one of the most important ways for improving the precision of the survey estimates. It is a method in which the designs weights are modified as little as possible by minimizing a given distance measure to the calibrated weights respecting a set of constraints related to suitable auxiliary information. This study proposes a new approach for Calibration Estimation (CE) of the population mean under stratified sampling scheme in the presence of two auxiliary variables. Almost all calibration estimation’s literature used Lagrange Multiplier Technique (LMT) in order to estimate the calibrated weights. While LMT requires all equations included in the model to be differentiable functions; some un- differentiable functions may be faced in some cases. Hence it is essential to look for using another technique that can provide more flexibility in dealing with the problem. Therefore, this study is concerned with using goal programming approach as a different approach for calibration estimation. This approach is conducted in two different cases; the first case that deals with Multivariate Calibration Estimation (MCE) approach for population mean of one study variable. While the second represents the Generalized Calibration Estimation (GCE) approach for population mean of two study variables.
A comparison study is conducted using actual and generated data to evaluate the performance of the proposed approach for multivariate calibration estimator with other existing calibration estimators. The results of the simulation study proves that using the proposed approach for multivariate calibration estimation is more flexible and efficient compared to other calibration estimation methods of the population mean.