The algorithm of noisy k-means.
作者： Sébastien Loustau , Camille Brunet
DOI:
关键词:
摘要: In this note, we introduce a new algorithm to deal with finite dimensional clustering errors in variables. The design of is based on recent theoretical advances (see Loustau (2013a,b)) statistical learning As the previous mentioned papers, mixes different tools from inverse problem literature and machine community. Coarsely, it two-step procedure: (1) deconvolution step noisy inputs (2) Newton's iterations as popular k-means.
archives-ouvertes.fr
harvard.edu
arxiv.org
128.84.21.199参考文章(11)
Alexander Meister, Deconvolution Problems in Nonparametric Statistics ,(2009)
Harald Luschgy, Siegfried Graf, Foundations of Quantization for Probability Distributions ,(2000)
Sébastien Loustau, Anisotropic oracle inequalities in noisy quantization arXiv: Statistics Theory. ,(2013)
Alexandre B. Tsybakov, Introduction à l'estimation non-paramétrique Springer. ,(2003)
Sébastien Loustau, Inverse statistical learning Electronic Journal of Statistics. ,vol. 7, pp. 2065- 2097 ,(2013) , 10.1214/13-EJS838
Sébastien Loustau, Michael Chichignoud, Adaptive Noisy Clustering arXiv: Statistics Theory. ,(2013)
David Pollard, A Central Limit Theorem for $k$-Means Clustering Annals of Probability. ,vol. 10, pp. 919- 926 ,(1982) , 10.1214/AOP/1176993713
M. P. Wand, Fast Computation of Multivariate Kernel Estimators Journal of Computational and Graphical Statistics. ,vol. 3, pp. 433- 445 ,(1994) , 10.1080/10618600.1994.10474656
L Cavalier, Nonparametric statistical inverse problems Inverse Problems. ,vol. 24, pp. 034004- ,(2008) , 10.1088/0266-5611/24/3/034004
John A. Hartigan, Clustering Algorithms ,(1975)