作者: Ke Liu , Yong-Qing Cheng , Jing-Yu Yang
DOI: 10.1016/0031-3203(92)90136-7
关键词: Applied mathematics 、 Algorithm 、 Moore–Penrose pseudoinverse 、 Feature extraction 、 Optimal discriminant analysis 、 Discriminant 、 Set (abstract data type) 、 Pattern recognition (psychology) 、 Linear discriminant analysis 、 Decomposition method (constraint satisfaction) 、 Mathematics
摘要: Abstract A generalized optimal set of discriminant vectors for linear feature extraction is presented. First, the criteria selecting are introduced, and then a unified solving method derived to solve in both cases large number samples small samples. The experimental results show that present superior Foley-Sammon (Foley Sammon, IEEE Trans. Comput. 24, 281–289 (1975)), positive pseudoinverse (Tian et al., Opt. Engng 25 (7), 834–839 (1986)), perturbation (Hong Yang, Pattern Recognition 24 , 317–324 (1991)), matrix rank decomposition (Cheng 101–111 (1992)) terms correct classification rate.