PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Sparse kernel orthonormalized PLS for feature extraction in large data sets
Jerónimo Arenas-Garcia, Kaare B. Petersen and Lars K. Hansen
In: Neural Information Processing Systems (NIPS'06), Vancouver(2007).


In this paper we are presenting a novel multivariate analysis method. Our scheme is based on a novel kernel orthonormalized partial least squares (PLS) variant for feature extraction, imposing sparsity constrains in the solution to improve scalability. The algorithm is tested on a benchmark of UCI data sets, and on the analysis of integrated short-time music features for genre prediction. The upshot is that the method has strong expressive power even with rather few features, is clearly outperforming the ordinary kernel PLS, and therefore is an appealing method for feature extraction of labelled data.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Theory & Algorithms
ID Code:5053
Deposited By:Jerónimo Arenas-Garcia
Deposited On:24 March 2009