TY - JOUR
T1 - Constructing descriptive and discriminative nonlinear features
T2 - Rayleigh coefficients in Kernel feature spaces
AU - Mika, Sebastian
AU - Rätsch, Gunnar
AU - Weston, Jason
AU - Schölkopf, Bernhard
AU - Smola, Alex
AU - Müller, Klaus Robert
PY - 2003/5
Y1 - 2003/5
N2 - We incorporate prior knowledge to construct nonlinear algorithms for invariant feature extraction and discrimination. Employing a unified framework in terms of a nonlinearized variant of the Raylelgh coefficient, we propose nonlinear generalizations of Fisher's discriminant and oriented PCA using support vector kernel functions. Extensive simulations show the utility of our approach.
AB - We incorporate prior knowledge to construct nonlinear algorithms for invariant feature extraction and discrimination. Employing a unified framework in terms of a nonlinearized variant of the Raylelgh coefficient, we propose nonlinear generalizations of Fisher's discriminant and oriented PCA using support vector kernel functions. Extensive simulations show the utility of our approach.
KW - Fisher's discriminant
KW - Kernel functions
KW - Nonlinear feature extraction
KW - Oriented PCA
KW - Raylelgh coefficient
KW - Support vector machine
UR - http://www.scopus.com/inward/record.url?scp=0038633559&partnerID=8YFLogxK
U2 - 10.1109/TPAMI.2003.1195996
DO - 10.1109/TPAMI.2003.1195996
M3 - Article
SN - 0162-8828
VL - 25
SP - 623
EP - 628
JO - IEEE Transactions on Pattern Analysis and Machine Intelligence
JF - IEEE Transactions on Pattern Analysis and Machine Intelligence
IS - 5
ER -