快速核有监督局部保留投影算法

张亮; 黄曙光; 郭浩

doi:10.3724/SP.J.1146.2010.01044

快速核有监督局部保留投影算法

doi: 10.3724/SP.J.1146.2010.01044

计量
- 文章访问数: 4007
- HTML全文浏览量: 124
- PDF下载量: 801
- 被引次数: 0
出版历程
- 收稿日期: 2010-09-25
- 修回日期: 2011-01-11
- 刊出日期: 2011-05-19

A Fast Kernel Supervised Locality Preserving Projection Algorithm

摘要

摘要: 为了提取样本中的非线性模式，保持其中的流形结构以及减少投影时间，该文提出了一种快速核有监督局部保留投影算法。该算法使用有监督聚类选择法选取训练集的一个子集进行子集核主成分分析，然后在子集核主成分分析形成的子空间中进行有监督局部保留投影。实验结果表明：相对于有监督局部保留投影算法以及现有的几种流行特征提取方法，新算法能够取得更高的识别率；相对于现有的核投影算法，新算法的投影速度更快。在有些数据集上，只要普通核投影十分之一左右的时间，就能达到相同甚至更高的识别率。
- 模式识别 /
- 特征提取 /
- 有监督局部保留投影 /
- 子集核主成分分析
Abstract: To extract nonlinear patterns, preserve the manifold structure, and reduce the projection time, a Fast Kernel Supervised Locality Preserving Projection (FKSLPP) algorithm is proposed. This new algorithm firstly selects a subset of the training set by supervised cluster selection algorithm to do Subset Kernel Principal Component Analysis (SKPCA), and then Supervised Locality Preserving Projection (SLPP) is performed in SKPCA subspace. Experiments results show that compared with SLPP and some other popular feature extraction algorithms, FKSLPP can get higher recognition rates; compared with kernel projection algorithms of state of art, FKSLPP is much faster. In some datasets, FKSLPP can get same or higher recognition rates while costs only one-tenth processing time of the common kernel projection algorithms.
- Pattern recognition /
- Feature extraction /
- Supervised locality preserving projection /
- Subset kernel principal component analysis

HTML全文

参考文献(1)

Loog M and Duin R P W. Linear dimensionality reduction via a heteroscedastic extension of LDA: the Chernoff criterion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2004, 26(6): 732-739.[2] Das K and Nenadic Z. Approximate information discriminant analysis: a computationally simple heteroscedastic feature extraction technique[J]. Pattern Recognition, 2008, 41(5): 1548-1557.[3] He Xiao-fei. Locality preserving projections[D]. [Ph.D. dissertation], The University of Chicago, 2005.[4] 申中华, 潘永惠, 王士同. 有监督的局部保留投影降维算法[J]. 模式识别与人工智能, 2008, 21(2): 233-239.Shen Zhong-hua, Pan Yong-hui, and Wang Shi-tong. A supervised locality preserving projection algorithm for dimensionality reduction[J]. Pattern Recognition and Artificial Intelligence, 2008, 21(2): 233-239.[5] 王晓明, 王士同. 广义的监督局部保留投影算法[J]. 电子与信息学报, 2009, 31(8): 1840-1845.Wang Xiao-ming and Wang Shi-tong. Generalized supervised locality preserving projection[J]. Journal of Electronics Information Technology, 2009, 31(8): 1840-1845.[6] Feng Gui-yu, Hu De-wen, Zhang D, and Zhou Zong-tan. An alternative formulation of kernel LPP with application to image recognition[J]. Neurocomputing, 2006, 69(13-15): 1733-1738.[7] Yang Jian, Jin Zhong, and Yang Jing-yu, et al.Essence of kernel Fisher discriminant: KPCA plus LDA[J]. Pattern Recognition, 2004, 37(10): 2097-2100.[8] Yoshikazu Washizawa. Subset kernel principal component analysis[C]. IEEE International Workshop on Machine Learning for Signal Processing, Grenoble, France, September 2-4, 2009: 1-6.[9] 段明秀. 层次聚类算法的研究及应用[D]. [博士论文], 中南大学, 2009.Duan Ming-Xiu. Research and application of hierarchical clustering algorithm[D]. [Ph.D. dissertation], Central South University, 2009.[10] Theodoridis S and Koutroumbas K. Pattern Recognition[M]. Fourth edition, Beijing: China Machine Press, 2009: 278-280.[11] Rueda L and Herrera M. Linear dimensionality reduction by maximizing the Chernoff distance in the transformed space[J]. Pattern Recognition, 2008, 41(10): 3138-3152.

施引文献

资源附件(0)

访问统计