Spectra Classification Based on Generalized Discriminant Analysis
XU Xin1，2, YANG Jin-fu1, WU Fu-chao1, ZHAO Yong-heng2
1. National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100080, China 2. National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012, China
Abstract：A kernel based generalized discriminant analysis (GDA) technique is proposed for the classification of stars, galaxies, and quasars. GDA combines the LDA algorithm with kernel trick, and samples are projected by nonlinear mapping onto the feature space F with high dimensions, and then LDA is conducted in F. Also, it could be inferred that GDA which combines the extension of Fisher’s criterion with kernel trick is complementary to kernel Fisher discriminant framework. LDA, GDA, PCA and KPCA were experimentally compared with these three different kinds of spectra. Among these four techniques, GDA obtains the best result, followed by LDA, and PCA is the worst. Although KPCA is also a kernel based technique, its performance is not satisfactory if the selected number of the principal components is small, and in some cases, it appears even worse than LDA, a non-kernel based technique.
Key words：Spectra classification;Generalized discriminant analysis;Linear discriminant analysis;Kernel principal component analysis
[1] QIN Dong-mei, HU Zhan-yi, ZHAO Yong-heng(覃冬梅, 胡占义, 赵永恒). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2003, 23(1): 182. [2] Cabanac R A, de Lapparent V, Hickson P. Astronomy and Astrophysics, 2002, 389: 1090. [3] Weaver W B, et al. Astrophysical Journal, 1997, 487: 847. [4] Vladimir N Vapnik. Translated by ZHANG Xue-gong(张学工译). Nature of Statistical Learning Theory(统计学习理论的本质). Beijing: Tsinghua University Press(北京: 清华大学出版社), 2000. [5] Scholkopf B, Smola A, Muller K R. Neural Computation, 1998, 10(5): 1299. [6] Mika S, Ratsch G, Weston J, et al. IEEE International Workshop Neural Networks for Signal Processing IX, 1999, (Aug.): 41. [7] Baudat G, Anouar F. Neural Computation, 2000, 12(10): 2385. [8] Richard O Duda, Peter E Hart, David G Stork. Transted by LI Hong-dong, YAO Tian-xiang, et al(李宏东, 姚天翔, 等译). Pattern Classification(Second Edition)(模式识别, 第2版). Beijing: China Machine Press(北京: 机械工业出版社), 2003. [9] Chen Li-fen, et al. Pattern Recognition, 2000, 33(10): 1713. [10] Yang Jian, Frangi A F, et al. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(2): 230. [11] Huang R, Liu Q S, Lu H Q, et al. Solving the Small Sample Size Problem of LDA. Proceedings of 16th International Conference on Pattern Recognition, 2002, 3: 29C32.
