Robust multivariate L1 principal component analysis and dimensionality reduction
作者： Junbin Gao , Paul W. Kwan , Yi Guo
DOI: 10.1016/J.NEUCOM.2008.01.027
关键词: Pattern recognition 、 Algorithm 、 Laplace distribution 、 Dimensionality reduction 、 Multivariate statistics 、 Sparse PCA 、 Mathematics 、 Principal component analysis 、 Expectation–maximization algorithm 、 Heavy-tailed distribution 、 Artificial intelligence 、 Bayesian inference
摘要: Further to our recent work on the robust L1 PCA we introduce a new version of model based so-called multivariate Laplace distribution (called distribution) proposed in Eltoft et al. [2006. On distribution. IEEE Signal Process. Lett. 13(5), 300-303]. Due heavy tail and high component dependency characteristics distribution, is expected be more against data outliers fitting dependency. Additionally, demonstrate how variational approximation scheme enables effective inference key parameters probabilistic L1-PCA model. By doing so, tractable Bayesian can achieved EM-type algorithm.
elsevier.com
sci-hub.se 参考文章(19)
Fernando De La Torre, Michael J Black, A Framework for Robust Subspace Learning International Journal of Computer Vision. ,vol. 54, pp. 117- 142 ,(2003) , 10.1023/A:1023709501986
Massimiliano Pontil, Sayan Mukherjee, Federico Girosi, On the Noise Model of Support Vector Machines Regression Lecture Notes in Computer Science. pp. 316- 324 ,(2000) , 10.1007/3-540-40992-0_24
Cédric Archambeau, Probabilistic models in noisy environments : and their application to a visual prosthesis for the blind/ ,(2005)
Donald B. Rubin, Iteratively Reweighted Least Squares Wiley StatsRef: Statistics Reference Online. ,(2014) , 10.1002/9781118445112.STAT03199
Michael E. Tipping, Neil D. Lawrence, Variational inference for Student-t models: Robust Bayesian interpolation and generalised component analysis Neurocomputing. ,vol. 69, pp. 123- 141 ,(2005) , 10.1016/J.NEUCOM.2005.02.016
Andrew Y. Ng, Feature selection, L1 vs. L2 regularization, and rotational invariance Twenty-first international conference on Machine learning - ICML '04. pp. 78- ,(2004) , 10.1145/1015330.1015435
D. de Ridder, V. Franc, Robust subspace mixture models using t-distributions. british machine vision conference. pp. 1- 10 ,(2003) , 10.5244/C.17.35
Y. Guo, J. Gao, P.W. Kwan, Twin Kernel Embedding IEEE Transactions on Pattern Analysis and Machine Intelligence. ,vol. 30, pp. 1490- 1495 ,(2008) , 10.1109/TPAMI.2008.74
Chris Ding, Ding Zhou, Xiaofeng He, Hongyuan Zha, R1-PCA Proceedings of the 23rd international conference on Machine learning - ICML '06. pp. 281- 288 ,(2006) , 10.1145/1143844.1143880
Qifa Ke, T. Kanade, Robust L/sub 1/ norm factorization in the presence of outliers and missing data by alternative convex programming computer vision and pattern recognition. ,vol. 1, pp. 739- 746 ,(2005) , 10.1109/CVPR.2005.309