ML Anthology

Subset Kernel PCA for Pattern Recognition

Yoshikazu Washizawa
ICCVW 2009 pp. 162-169
doi:10.1109/ICCVW.2009.5457705 /iccvw/2009/washizawa2009iccvw-subset/

Abstract

Subspace methods that utilize principal component analysis (PCA) are widely used for pattern classification or detection problems. Kernel PCA (KPCA) that is an extension of PCA is also applied to subspace methods. However, its computational cost is very high since the computational cost mainly depends on the number of samples in kernel methods. Recently, subset KPCA (SKPCA) has been proposed in order to reduce its computational complexity. In this paper, we apply SKPCA to subspace methods, and compare SKPCA with KPCA using some sample selection methods. Experimental results demonstrate advantages of subspace methods using SKPCA.

ICCVW Semantic Scholar

Cite

Text

Washizawa. "Subset Kernel PCA for Pattern Recognition." IEEE/CVF International Conference on Computer Vision Workshops, 2009. doi:10.1109/ICCVW.2009.5457705

Markdown

[Washizawa. "Subset Kernel PCA for Pattern Recognition." IEEE/CVF International Conference on Computer Vision Workshops, 2009.](https://mlanthology.org/iccvw/2009/washizawa2009iccvw-subset/) doi:10.1109/ICCVW.2009.5457705

BibTeX

@inproceedings{washizawa2009iccvw-subset,
  title     = {{Subset Kernel PCA for Pattern Recognition}},
  author    = {Washizawa, Yoshikazu},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2009},
  pages     = {162-169},
  doi       = {10.1109/ICCVW.2009.5457705},
  url       = {https://mlanthology.org/iccvw/2009/washizawa2009iccvw-subset/}
}