%L eprints1761
%A Giorgio Gnecco
%A Marcello Sanguineti
%V 4
%N 2
%D 2010
%I Springer
%R 10.1007/s11590-009-0158-1
%K Principal component analysis (PCA); Kernel methods; Suboptimal solutions
%P 197-210
%X Suboptimal solutions to kernel principal component analysis are considered. Such solutions take on the form of linear combinations of all n-tuples 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.
%T Error Bounds for Suboptimal Solutions to Kernel Principal Component Analysis
%J Optimization Letters