Title :
Optimal Feature Representation for Kernel Machines using Kernel-Target Alignment Criterion
Author :
Pothin, Jean-Baptiste ; Richard, Cédric
Author_Institution :
Institut Charles Delaunay, Univ. de Technol. de Troyes
Abstract :
Kernel-target alignment is commonly used to predict the behavior of any given reproducing kernel in a classification context, without training any kernel machine. In this paper, we present a gradient ascent algorithm for maximizing the alignment over linear transform of the input space. Our method is compared to the minimization of the radius-margin bound. Experimental results on multi-dimensional benchmarks show the effectiveness of our approach
Keywords :
data handling; gradient methods; minimisation; transforms; gradient ascent algorithm; kernel machines; kernel-target alignment criterion; linear transform; multi-dimensional benchmarks; optimal feature representation; radius-margin bound minimization; Algorithm design and analysis; Design optimization; Euclidean distance; Extraterrestrial measurements; Kernel; Large-scale systems; Minimization methods; Pattern classification; Support vector machine classification; Support vector machines; SVM; alignment; pattern classification;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0727-3
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2007.366867
