|
Minimum-covariance-determinant-based mean estimators under systematic sampling |
|---|---|
| รหัสดีโอไอ | |
| Creator | Usman Shahzad |
| Title | Minimum-covariance-determinant-based mean estimators under systematic sampling |
| Publisher | Maejo University |
| Publication Year | 2568 |
| Journal Title | Maejo International Journal of Science and Technology |
| Journal Vol. | 19 |
| Journal No. | 1 |
| Page no. | 67 |
| Keyword | minimum covariance determinant, mean estimation, robust regression, systematic random sampling |
| Website title | Maejo International Journal of Science and Technology |
| ISSN | 1905-7873 |
| Abstract | The ordinary least square (OLS) estimates become inappropriate in the presence of outliers and consequently the mean estimators based on the OLS coefficients also become unsuitable. To address this issue, employing robust regression tools and covariance matrices for mean estimation is a well-established practice under a simple random sampling scheme. However, the mean estimation using robust regression tools and covariance matrices under systematic sampling has not been explored yet. To fill this gap, we develop a class of mean estimators under systematic sampling in this article. This study proposes a family of minimum-covariance-determinant-based mean estimators along with their theoretical properties. The Dixon test is used to con?rm the presence of extremes in the data. Numerical analyses related to real and simulated data are performed to assess the new proposals. According to the percentage of relative efficiency in the practical study with timber volume data, estimators based on the Huber regression show a percentage relative efficiency (PRE) increase from 226.16 to 241.68. In simulated data scenarios, the PRE of various robust estimators exceeds 450. Hence the proposed estimators provide excellent performance. |
