%0 Journal Article
%T Efficient Kernel Principal Component Analysis Algorithm for Large-Scale Data Set<br>一种解决大规模数据集问题的核主成分分析算法
%A SHI Wei-Ya
%A GUO Yue-Fei
%A XUE Xiang-Yang
%A <br>史卫亚
%A 郭跃飞
%A 薛向阳
%J 软件学报
%D 2009
%I 
%X A covariance-free method of computing kernel principal components is proposed. First, a matrix, called Gram-power matrix, is constructed with the original Gram matrix. It is proven by the theorem of linear algebra that the eigenvectors of newly constructed matrix are the same as those of the Gram matrix. Therefore, each column of the Gram matrix can be treated as the input sample for the iterative algorithm. Thus, the kernel principle components can be iteratively computed without the eigen-decomposition. The space complexity of the proposed method is only O(m), and the time complexity is reduced to O(pkm). The effectiveness of the proposed method is validated by experimental results. More importantly, it still can be used even if traditional eigen-decomposition technique cannot be applied when faced with the extremely large-scale data set.
%K KPCA (kernel principal component analysis)
%K Gram matrix
%K large-scale data set
%K covariance-free
%K eigen-decomposition<br>核主成分分析
%K Gram矩阵
%K 大规模数据集
%K 协方差无关
%K 特征分解
%U http://www.alljournals.cn/get_abstract_url.aspx?pcid=5B3AB970F71A803DEACDC0559115BFCF0A068CD97DD29835&cid=8240383F08CE46C8B05036380D75B607&jid=7735F413D429542E610B3D6AC0D5EC59&aid=A1A2A08609B9E3071739F30F3F589F50&yid=DE12191FBD62783C&vid=A04140E723CB732E&iid=5D311CA918CA9A03&sid=8BA849EBE34B1882&eid=37DABCB9D67C5C3E&journal_id=1000-9825&journal_name=软件学报&referenced_num=1&reference_num=13