Small Sample Optimal Discriminant Transform Based on PSO under Fisher Criterion
RUI Ting1, ZHOU You2, QI Tian3, FANG Hu-Sheng1, RONG Xiao-Li1
1.Engineering Institute of Engineering Corps, PLA University of Science and Technology, Nanjing 210007 2.Accounting Department, Jiangsu Institute of Economic and Trade Technology, Nanjing 210007 3.Department of Computer Science, University of Texas at San Antonio, San Antonio, TX 78249
Abstract The within-class scatter matrix Fisher criterion is singular under small samples. Therefore, it can not be solved directly. A method based on PSO is proposed to get optimal discriminant transform under small samples without calculating inverse of the within-class scatter matrix. The methods and steps are discussed to get optimal discriminant projection vector by velocity-position search model of particle swarm optimization. The eigenvectors method and the proposed method are compared, when within-class scatter matrix is non-singular. Experimental results on both small and large samples demonstrate the accuracy of the proposed method.
[1] Tian Q, Fainman Y, Gu Z H, et al. Comparison of Statistical Pattern Recognition Algorithms for Hybrid Processing, Part II: Eigenvector-Based Algorithms. Journal of the Optical Society of America, 1988, 5(10): 1670-1672 [2] Hong Ziquan,Yang Jingyu. Optimal Discriminant Plane for a Small Number of Samples and Design Method of Classifier on the Plane. Pattern Recognition, 1991, 24(4): 314-324 [3] Cheng Yongqing, Zhuang Yongming, Yang Jingyu. Optimal Fisher Discriminant Analysis Using the Rank Decomposition. Pattern Recognition, 1992, 25(1): 101-111 [4] Kennedy J, Eberhart R C. Particle Swarm Optimization // Proc of the IEEE International Conference on Neural Networks. Perth Western, Australia, 1995, Ⅳ: 1942-1946 [5] Zitzler E, Deb K, Thiele L. Comparison of Multiobjective Evolutionary Algorithms: Empirical Results. Evolutionary Computation, 2000, 8(2): 173-195 [6] Ciuprina G, Laon D, Munteanu I. Use of Intelligent-Particle Swarm Optimization in Electromagnetics. IEEE Trans on Magnetics, 2002, 38(2): 1037-1040 [7] Mendes R, Kennedy J, Neves J. The Fully Informed Particle Swarm: Simpler, Maybe Better. IEEE Trans on Evolutionary Computation, 2004, 8(3): 204-210 [8] Liang J J, Qin A K. Comprehensive Learning Particle Swarm Optimizer for Global Optimization of Multi-Modal Functions. IEEE Trans on Evolutionary Computation, 2006, 10(3): 281-295 [9] Lü Yanping, Li Shaozi, Chen Shuili, et al. Particle Swarm Optimization Based on Adaptive Diffusion and Hybrid Mutation. Journal of Software, 2007, 18(11): 2740-2751 (in Chinese) (吕艳萍,李绍滋,陈水利,等.自适应扩散混合变异机制微粒群算法.软件学报, 2007, 18(11): 2740-2751) [10] Rui Ting, Shen Chunlin, Qi Tian, et al. Comparison and Analysis on ICA & PCA's Ability in Feature Extraction. Pattern Recognition and Artificial Intelligence, 2005, 18(1): 124-128 (in Chinese) (芮 挺,沈春林,Qi Tian,等.ICA与PCA特征抽取能力的比较分析.模式识别与人工智能, 2005, 18(1): 124-128) [11] Cortes C, Vapnik V. Support Vector Networks. Machine Learning, 1995, 20(3): 273-295
2009, Vol. 22 
