مشروع البحث: Reducing the Effect of Outliers in Nonlinear Regression Model Using Robust Estimatiors
| dc.contributor.advisor | السعيدي المهدي محمد الطاهر | |
| dc.date.accessioned | 2025-06-23T09:29:46Z | |
| dc.date.available | 2025-06-23T09:29:46Z | |
| dc.description | Nonlinear least squares (NLS) method is one of the most commonly used for estimating the parameters of nonlinear regression models. In the presence of outliers, even one unusual observation may have a significant impact on the final estimates, and for this it has been suggested the use of robust methods. This study aims to identify the most common robust nonlinear techniques that are used as a better alternative to classical least squares. This includes M estimator, MM estimator, CM estimator, TAU estimator and MTL estimator. In addition, the goal is to compare their practical performance under different conditions (sample size, percentage of outliers and model formula).Simulation results (using the programming language R) indicated that the best performance was achieved by the MM, M and TAU estimator for all percentage of outliers (10%, 20%, 40%), all sample sizes (50, 100, 150) as well as all five regression models. To confirm the effectiveness of these methods in practice, real data sets have been used with outliers, consisting of Drug Concentration data and Cow Milk data. Through the results of the application of these estimators we found that the best of these estimators is TAU estimator. Furthermore, the results confirmed that the NLS estimator remains the best when there are no outliers. | |
| dc.description.abstract | Nonlinear least squares (NLS) method is one of the most commonly used for estimating the parameters of nonlinear regression models. In the presence of outliers, even one unusual observation may have a significant impact on the final estimates, and for this it has been suggested the use of robust methods. This study aims to identify the most common robust nonlinear techniques that are used as a better alternative to classical least squares. This includes M estimator, MM estimator, CM estimator, TAU estimator and MTL estimator. In addition, the goal is to compare their practical performance under different conditions (sample size, percentage of outliers and model formula).Simulation results (using the programming language R) indicated that the best performance was achieved by the MM, M and TAU estimator for all percentage of outliers (10%, 20%, 40%), all sample sizes (50, 100, 150) as well as all five regression models. To confirm the effectiveness of these methods in practice, real data sets have been used with outliers, consisting of Drug Concentration data and Cow Milk data. Through the results of the application of these estimators we found that the best of these estimators is TAU estimator. Furthermore, the results confirmed that the NLS estimator remains the best when there are no outliers. | |
| dc.identifier | 7201 | |
| dc.identifier.uri | https://dspace.academy.edu.ly/handle/123456789/1649 | |
| dc.subject | Reducing the Effect of Outliers in Nonlinear Regression Model Using Robust Estimatiors | |
| dc.title | Reducing the Effect of Outliers in Nonlinear Regression Model Using Robust Estimatiors | |
| dspace.entity.type | Project | |
| project.endDate | 2020 | |
| project.funder.name | رياضيات | |
| project.investigator | عبد القادر خليفة امحمد السالم | |
| project.startDate | 2019 | |
| relation.isOrgUnitOfProject | abd0d318-9280-4b6c-a480-6db4a3a7c674 | |
| relation.isOrgUnitOfProject.latestForDiscovery | abd0d318-9280-4b6c-a480-6db4a3a7c674 |