Advances and Applications in Statistics
Volume 52, Issue 6, Pages 431 - 447
(June 2018) http://dx.doi.org/10.17654/AS052060431 |
|
WILCOXON RANK BASED PRINCIPAL COMPONENT ANALYSIS
Hend A. Auda, Mohamed A. Ismail and Ali A. Rashed
|
Abstract: A novel method of principal component analysis (PCA) is proposed. The L2-norm in the traditional PCA is substituted by the Wilcoxon rank based norm. Brooks et al. [3] introduced the PCA based on L1-norm to counter the effect of outliers and non-normality. A simulation study is conducted to compare the performance of the Wilcoxon rank based PCA with the L2-PCA and L1-PCA. The simulation results show the outperformance of the proposed method.
Job quality index is computed. The first PC explained 29%, 90% and 96:5% of the total variation using L2, L1 and Wilcoxon rank based norm, respectively. |
Keywords and phrases: non-parametric PCA, L1-norm PCA, Wilcoxon rank based norm, job quality index. |
|
Number of Downloads: 443 | Number of Views: 3555 |
|