Unsupervised Classification with Non-Gaussian Mixture Models Using ICA
作者： Terrence J. Sejnowski , Michael S. Lewicki , Te-Won Lee
DOI:
关键词:
摘要: We present an unsupervised classification algorithm based on ICA mixture model. The model assumes that the observed data can be categorized into several mutually exclusive classes in which components each class are generated by a linear of independent sources. finds sources, mixing matrix for and also computes membership probability point. This approach extends Gaussian so have non-Gaussian structure. demonstrate this method learn efficient codes to represent images natural scenes text. learned basis functions yield better approximation underlying distributions data, thus provide greater coding efficiency. believe is well suited modeling structure high-dimensional has many potential applications.
nips.cc
acm.org 参考文章(12)
Bruno A. Olshausen, Michael S. Lewicki, Inferring sparse, overcomplete image codes using an efficient coding framework neural information processing systems. pp. 815- 821 ,(1998)
Z Ghahramani, Solving inverse problems using an EM approach to density estimation Erlbaum Associates. ,(1993)
John Stutz, Peter Cheeseman, None, Autoclass — A Bayesian Approach to Classification Springer, Dordrecht. pp. 117- 126 ,(1996) , 10.1007/978-94-009-0107-0_13
Christopher M. Bishop, Mixture density networks Aston University. ,(1994)
J.-F. Cardoso, B.H. Laheld, Equivariant adaptive source separation IEEE Transactions on Signal Processing. ,vol. 44, pp. 3017- 3030 ,(1996) , 10.1109/78.553476
Te-Won Lee, Mark Girolami, Terrence J. Sejnowski, Independent component analysis using an extended infomax algorithm for mixed subgaussian and supergaussian sources Neural Computation. ,vol. 11, pp. 417- 441 ,(1999) , 10.1162/089976699300016719
Anthony J. Bell, Terrence J. Sejnowski, An information-maximization approach to blind separation and blind deconvolution Neural Computation. ,vol. 7, pp. 1129- 1159 ,(1995) , 10.1162/NECO.1995.7.6.1129
Mark A. Girolami, An Alternative Perspective on Adaptive Independent Component Analysis Algorithms Neural Computation. ,vol. 10, pp. 2103- 2114 ,(1998) , 10.1162/089976698300016981
Anthony J. Bell, Terrence J. Sejnowski, The "independent components" of natural scenes are edge filters. Vision Research. ,vol. 37, pp. 3327- 3338 ,(1997) , 10.1016/S0042-6989(97)00121-1
Bruno A. Olshausen, David J. Field, Emergence of simple-cell receptive field properties by learning a sparse code for natural images Nature. ,vol. 381, pp. 607- 609 ,(1996) , 10.1038/381607A0