Abstract
Classic kernel principal component analysis (KPCA) is less computationally efficient when extracting features from large data sets. In this paper, we propose an algorithm, that is, efficient KPCA (EKPCA), that enhances the computational efficiency of KPCA by using a linear combination of a small portion of training samples, referred to as basic patterns, to approximately express the KPCA feature extractor, that is, the eigenvector of the covariance matrix in the feature extraction. We show that the feature correlation (i.e., the correlation between different feature components) can be evaluated by the cosine distance between the kernel vectors, which are the column vectors in the kernel matrix. The proposed algorithm can be easily implemented. It first uses feature correlation evaluation to determine the basic patterns and then uses these to reconstruct the KPCA model, perform feature extraction, and classify the test samples. Since there are usually many fewer basic patterns than training samples, EKPCA feature extraction is much more computationally efficient than that of KPCA. Experimental results on several benchmark data sets show that EKPCA is much faster than KPCA while achieving similar classification performance.
Similar content being viewed by others
References
Schölkopf B, Smola A, Muller KR (1998) Nonlinear component analysis as a kernel eigenvalue problem. Neural Comput 10(5):1299–1319
Muller KR, Mika S, Ratsch G, Tsuda K et al (2001) An introduction to kernel-based learning algorithms. IEEE Trans Neural Netw 12(2):181–201
Shawe-Taylor J, Cristianini N (2004) Kernel methods for pattern analysis. Cambridge University Press, Cambridge
Scholkopf B, Mika S, Burges CJC, Knirsch P et al (1999) Input space versus feature space in kernel-based methods. IEEE Trans Neural Netw 10(5):1000–1017
Honeine P (2012) Online kernel principal component analysis: a reduced-order model. IEEE Trans Pattern Anal Mach Intell 34(9):1814–1826
Abrahamsen TJ, Hansen LK (2011) A cure for variance inflation in high dimensional kernel principal component analysis. J Mach Learn Res 12(6):2027–2044
Taouali O, Elaissi I, Messaoud H (2012) Online identification of nonlinear system using reduced kernel principal component analysis. Neural Comput Appl 21:161–169
Rasmussen PM, Abrahamsen TJ, Madsen KH, Hansen LK (2012) Nonlinear denoising and analysis of neuroimages with kernel principal component analysis and pre-image estimation. NeuroImage 60(3):1807–1818
Honeine P, Richard C (2011) Preimage problem in kernel-based machine learning. Signal Process Mag IEEE 28(2):77–88
Mika S, Schölkopf B, Smola A, Müller KR et al (1999) Kernel PCA and de-noising in feature spaces. Adv Neural Inform Process Syst 11(1):536–542
Zheng WS, Lai JH, Yuen PC (2010) Penalized preimage learning in kernel principal component analysis. IEEE Trans Neural Netw 21(4):551–570
Kim KI, Franz MO, Scholkopf B (2005) Iterative kernel principal component analysis for image modeling. IEEE Trans Pattern Anal Mach Intell 27(9):1351–1366
Liu C (2004) Gabor-based kernel PCA with fractional power polynomial models for face recognition. IEEE Trans Pattern Anal Mach Intell 26(5):572–581
De la Torre F, Nguyen MH (2008) Parameterized kernel principal component analysis: Theory and applications to supervised and unsupervised image alignment. In: IEEE conference on computer vision and pattern recognition, 2008. CVPR 2008, pp. 1–8
Li J, Li X, Tao D (2008) KPCA for semantic object extraction in images. Pattern Recogn 41(10):3244–3250
Hotta K (2012) Local co-occurrence features in subspace obtained by KPCA of local blob visual words for scene classification. Pattern Recogn 45(10):3687–3694
Turk M, Pentland A (1991) Eigenfaces for recognition. J Cogn Neurosci 3(1):71–86
Yang J, Zhang D, Frangi AF (2004) Two-dimensional PCA: a new approach to appearance-based face representation and recognition. IEEE Trans Pattern Anal Mach Intell 26(1):131–137
Zhang D, Zhou ZH (2005) (2D) 2PCA: two-directional two-dimensional PCA for efficient face representation and recognition. Neurocomputing 69(1):224–231
Xu Y, Zhang D, Yang JY (2010) A feature extraction method for use with bimodal biometrics. Pattern Recogn 43(3):1106–1115
Vapnik V (1999) The nature of statistical learning theory. Springer, Berlin
Boser BE, Guyon IM, Vapnik VN (1992) A training algorithm for optimal margin classifiers. In: Proceedings of the fifth annual workshop on Computational learning theory, pp 144–152
Lee D, Jung KH, Lee J (2009) Constructing sparse kernel machines using attractors. IEEE Trans Neural Netw 20(4):721–729
Rosipal R, Girolami M (2001) An expectation-maximization approach to nonlinear component analysis. Neural Comput 13(3):505–510
Zheng W, Zou C, Zhao L (2005) An improved algorithm for kernel principal component analysis. Neural Process Lett 22(1):49–56
Schraudolph NN, Gunter S, Vishwanathan S (2007) Fast iterative kernel PCA. Presented at the advances in neural information processing systems
Moerland P (2000) An on-line EM algorithm applied to kernel PCA, Technical report, IDIAP2000
Gnecco G, Sanguineti M (2009) Accuracy of suboptimal solutions to kernel principal component analysis. Comput Optim Appl 42(2):265–287
Xu Y, Zhang D, Song F, Yang JY et al (2007) A method for speeding up feature extraction based on KPCA. Neurocomputing 70(4):1056–1061
Kim SW, Oommen BJ (2004) On using prototype reduction schemes to optimize kernel-based nonlinear subspace methods. Pattern Recogn 37(2):227–239
Chin TJ, Suter D (2007) Incremental kernel principal component analysis. IEEE Trans Image Process 16(6):1662–1674
Alzate C, Suykens J (2008) Kernel component analysis using an epsilon-insensitive robust loss function. IEEE Trans Neural Netw 19(9):1583–1598
Xu Y, Lin C, Zhao W (2010) Producing computationally efficient KPCA-based feature extraction for classification problems. Electron Lett 46(6):452–453
Wright J, Yang AY, Ganesh A, Sastry SS et al (2009) Robust face recognition via sparse representation. IEEE Trans Pattern Anal Mach Intell 31(2):210–227
Wang S-J, Yang J, Sun M-F, Peng X-J et al (2012) Sparse tensor discriminant color space for face verification. Neural Netw Learn Syst IEEE Trans 23(6):876–888
Wang S, Yang J, Zhang N, Zhou C (2011) Tensor discriminant color space for face recognition. IEEE Trans Image Process 20(9):2490–2501
Wang S-J, Zhou C-G, Zhang N, Peng X-J et al (2011) Face recognition using second-order discriminant tensor subspace analysis. Neurocomputing 74(12):2142–2156
Wang S-J, Chen H-L, Peng X-J, Zhou C-G (2011) Exponential locality preserving projections for small sample size problem. Neurocomputing 74(17):3654–3662
Ishii T, Ashihara M, Abe S (2008) Kernel discriminant analysis based feature selection. Neurocomputing 71(13):2544–2552
Kim JS, Scott CD (2010) L2 kernel classification. IEEE Trans Pattern Anal Mach Intell 32(10):1822–1831
Tian J, Li M, Chen F (2009) A hybrid classification algorithm based on coevolutionary EBFNN and domain covering method. Neural Comput Appl 18(3):293–308
Vert JP, Tsuda K, Schölkopf B (2004) A primer on kernel methods. In: Schölkopf B, Tsuda K, Vert JP (eds) Kernel methods in computational biology, MIT Press, Cambridge, pp 35–70. http://www.amazon.com/Kernel-Methods-Computational-Biology-Molecular/dp/0262195097#reader_0262195097
Yang ZR (2006) A novel radial basis function neural network for discriminant analysis. IEEE Trans Neural Netw 17(3):604–612
Xu Y, Zhu Q, Wang J (2011) Breast cancer diagnosis based on a kernel orthogonal transform. Neural Comput Appl 21(8):1865–1870
Acknowledgments
This article is partly supported by NSFC under grants Nos. 61263032, 61071179, 11061014, and 61202276, Jiangxi Provincial Natural Science Foundation of China under Grant 2010GQS0027, as well as the Science and Technology Foundation of Jiangxi Educational Committee of China (GJJ12309).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Fan, Z., Wang, J., Xu, B. et al. An efficient KPCA algorithm based on feature correlation evaluation. Neural Comput & Applic 24, 1795–1806 (2014). https://doi.org/10.1007/s00521-013-1424-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-013-1424-9