Title :
A comparative study of linear and nonlinear feature extraction methods
Author :
Park, Cheong Hee ; Park, Haesun ; Pardalos, Panos
Author_Institution :
Dept. of Comput. Sci. & Eng., Minnesota Univ., Minneapolis, MN, USA
Abstract :
This paper presents theoretical relationships among several generalized LDA algorithms and proposes computationally efficient approaches for them utilizing the relationships. Generalized LDA algorithms are extended nonlinearly by kernel methods resulting in nonlinear discriminant analysis. Performances and computational complexities of these linear and nonlinear discriminant analysis algorithms are compared.
Keywords :
feature extraction; matrix algebra; computational complexity; generalized LDA algorithm; kernel method; linear discriminant analysis; linear feature extraction; nonlinear discriminant analysis; nonlinear feature extraction; Algorithm design and analysis; Chromium; Computational complexity; Eigenvalues and eigenfunctions; Feature extraction; Kernel; Linear discriminant analysis; Performance analysis; Scattering; Singular value decomposition;
Conference_Titel :
Data Mining, 2004. ICDM '04. Fourth IEEE International Conference on
Print_ISBN :
0-7695-2142-8
DOI :
10.1109/ICDM.2004.10066
