核矩阵逼近的贪心算法-模式识别与人工智能.pdfVIP

第20 卷摇 第 1期摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 模式识别与人工智能摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Vol.20摇 No.1 摇 2007年2月摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 PR AI摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 Feb摇 摇 2007 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 摇 核核矩阵逼近的贪心算法* 杜京义摇 摇 侯媛彬 (西安科技大学 电气与控制工程学院西安摇 710054) 摘摇 要摇 在综合考虑数据之间的相关性与残差范数最小的基础上,本文分别提出向前贪心算法、向后贪心算法和 混合贪心算法寻踪最佳低秩逼近.为此提出一种稀疏回归算法(SRA).SRA能有效减少“训练样本冶,并具备良好的 推广能力.将SRA应用于2 个实际的模式识别问题,并与支持向量机(SVM)、核主元回归(KPCR)和关键算法 (KA)进行比较,验证SRA 的有效性. 关键词摇 核算法,核矩阵,低秩矩阵,贪心算法 中图法分类号摇 TP301 Greedy Algorithms with Kernel Matrix Approximation DUJing蛳Yi,HOU Yuan蛳Bin (College of Electrical and Control Engineering,Xi爷an University of Science and Technology,Xi爷an710054) ABSTRACT Kernel algorithm is a new nonlinear technique in the machine learning field with great practical value . In the Kernel algorithm Kernel Matrix serves as a bridge between data 蛳in and learning algorithms. The low 蛳rank approximation of Kernel Matrix is an effective means of advancing the computational efficiency of Kernel algorithm and reducing the internal memory. Based on the correlation between data and minimal residual norm,the forward greedy algorithm,the backward greedy algorithmandthemixedgreedy algorithmareproposedtolookfortheoptimumlow蛳rankapproximation. A sparse regression algorithm (SRA) is presented which effectively cuts off “training samples冶 and keeps good generalization capacity. With comparisonto SVM,KA and KPCR,the superiority of SRA isdemon鄄 strated on two datasets. Key Words摇 Kernel Algorithm,Kernel Matrix,Low蛳Rank Matrix,Greedy Algorithm *陕西省科学基金资助项目(No.2004JC12) 收稿日期:2006蛳03蛳20;