Gnecco, Giorgio and Sanguineti, Marcello Error Bounds for Suboptimal Solutions to Kernel Principal Component Analysis. Optimization Letters, 4 (2). pp. 197210. ISSN 18624472 (2010)
This is the latest version of this item.
Abstract
Suboptimal solutions to kernel principal component analysis are considered. Such solutions take on the form of linear combinations of all ntuples of kernel functions centered on the data, where n is a positive integer smaller than the cardinality m of the data sample. Their accuracy in approximating the optimal solution, obtained in general for n = m, is estimated. The analysis made in Gnecco and Sanguineti (Comput Optim Appl 42:265–287, 2009) is extended. The estimates derived therein for the approximation of the first principal axis are improved and extensions to the successive principal axes are derived.
Item Type:  Article 

Identification Number:  10.1007/s1159000901581 
Uncontrolled Keywords:  Principal component analysis (PCA); Kernel methods; Suboptimal solutions 
Subjects:  Q Science > QA Mathematics > QA75 Electronic computers. Computer science 
Research Area:  Computer Science and Applications 
Depositing User:  Giorgio Gnecco 
Date Deposited:  17 Sep 2013 13:09 
Last Modified:  17 Sep 2013 13:09 
URI:  http://eprints.imtlucca.it/id/eprint/1761 
Available Versions of this Item

Error bounds for suboptimal solutions to kernel principal component analysis. (deposited 13 Sep 2013 09:40)
 Error Bounds for Suboptimal Solutions to Kernel Principal Component Analysis. (deposited 17 Sep 2013 13:09) [Currently Displayed]
Actions (login required)
Edit Item 