Comparative Performance Analysis of Three Algorithms for Principal Component Analysis
Principal Component Analysis (PCA) is an important concept in statistical signal processing. In this paper, we evaluate an on-line algorithm for PCA, which we denote as the Exact Eigendecomposition (EE) algorithm. The algorithm is evaluated using Monte Carlo Simulations and compared with the PAST an...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Spolecnost pro radioelektronicke inzenyrstvi
2006-12-01
|
| Series: | Radioengineering |
| Online Access: | http://www.radioeng.cz/fulltexts/2006/06_04_84_90.pdf |
| Summary: | Principal Component Analysis (PCA) is an important concept in statistical signal processing. In this paper, we evaluate an on-line algorithm for PCA, which we denote as the Exact Eigendecomposition (EE) algorithm. The algorithm is evaluated using Monte Carlo Simulations and compared with the PAST and RP algorithms. In addition, we investigate a normalization procedure of the eigenvectors for PAST and RP. The results show that EE has the best performance and that normalization improves the performance of PAST and RP algorithms, respectively. |
|---|---|
| ISSN: | 1210-2512 |