基于聚类的核矩阵维度缩减  

Clustering-Based Dimensionality Reduction for Kernel Matrix

作　　者：陈才扣 高林[1] 高秀梅[1] 杨静宇[1] 

机构地区：[1]南京理工大学计算机科学系

出　　处：《数据采集与处理》2004年第3期250-253,共4页Journal of Data Acquisition and Processing

基　　金：国家自然科学基金 ( 60 0 72 0 3 4)资助项目 ;高等学校博士点基金 ( 2 0 0 2 0 2 880 1 3 )资助项目

摘　　要：提出了基于聚类的核矩阵维度缩减技术。它的主要思想是首先利用非线性映射 将原始输入空间变换到某高维特征空间 ,然后根据 k-均值聚类算法缩减训练样本的数目 ,得到一缩减的代表集 ,利用该代表集计算得到一组标准正交的基向量 ,构成一个低维的投影子空间。CENPARMI手写体阿拉伯数字库的试验结果证实了所提算法的有效性。A novel method for dimensionality reduction of kernel matrix is presented. Its main idea is that the original input space is first transformed to a high dimensional feature space via a nonlinear mapping . The k-means clustering algorithm is used to reduce the number of the transformed training samples and a reduced set, also called representative set, is derived. Based on the representative set, a set of orthogonormal basis vectors is obtained to form a new lower dimensional projection subspace. The method is tested in CENPARMI handwriting digit database at Concordia University. Experimental results show the method can significantly reduce complexity of the found classifiers while retaining their accuracy.

关 键 词：聚类 核矩阵维度缩减 模式识别 KFDA 鉴别分析 特征空间 特征抽取算法 图像特征 手写体数字识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]