| 多重核线性判别分析及其权值优化 |
| |
| 引用本文： | 刘笑嶂,冯国灿. 多重核线性判别分析及其权值优化[J]. 计算机应用, 2009, 29(9) |
| |
| 作者姓名： | 刘笑嶂 冯国灿 |
| |
| 作者单位： | 河源职业技术学院,电子与信息工程系,广东,河源,517000;中山大学,数学与计算科学学院,广州,510275 |
| |
| 基金项目： | 国家自然科学基金资助项目，教育部重点基金资助项目 |
| |
| 摘 要： | 为了提高非线性分类精度,借鉴在支持向量机(SVM)框架下发展起来的多重核学习方法,针对基于核的线性判别分析(KLDA)构造多重核.进而,使用拉格朗日乘子法优化最大边缘准则(MMC),提出了多重核权值优化算法.在FERET和CMU PIE人脸图像库上的实验表明,与基于单个核的LDA相比,多重核线性判别分析能够达到更高的分类性能.
|
| 关 键 词： | 多重核 核线性判别分析 最大边缘准则 权值优化 拉格朗日乘子法 |
Multiple kernel discriminant analysis with optimized weight |
| |
| LIU Xiao-zhang,FENG Guo-can. Multiple kernel discriminant analysis with optimized weight[J]. Journal of Computer Applications, 2009, 29(9) |
| |
| Authors: | LIU Xiao-zhang FENG Guo-can |
| |
| Affiliation: | 1.Department of Electronics and Information Engineering;Heyuan Polytechnic;Heyuan Guangdong 517000;China;2.School of Mathematics and Computational Science;Sun Yat-sen University;Guangzhou Guangdong 510275;China |
| |
| Abstract: | In order to enhance the accuracy of nonlinear classification,the multiple kernel learning method developed under the framework of Support Vector Machine(SVM)was referred to.The authors constructed a multi-kernel for kernel-based Linear Discriminant Analysis(LDA).Moreover,a weight optimization scheme for the multi-kernel was proposed by maximizing the Margin Maximization Criterion(MMC)based on the method of Lagrange multipliers.The experiments on the FERET and CMU PIE face database show that multiple kernel ... |
| |
| Keywords: | |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
