An Efficient K -Hyperplane Clustering Algorithm and Its Application to Sparse Component Analysis

Based on eigenvalue decomposition, a novel efficient K-HPC algorithm is developed in this paper, which is easy to implement. And it enables us to detect the number of hyperplanes and helps to avoid local minima by overestimating the number of hyperplanes. A confidence index is proposed to evaluate which estimated hyperplanes are most significant and which are spurious. So we can choose those significant hyperplanes with high rank priority and remove the spurious hyperplanes according to their corresponding confidence indices. Furthermore, a multilayer clustering framework called "multilayer K-HPC" is proposed to further improve the clustering results. The K-HPC approach can be directly applied to sparse component analysis (SCA) to develop efficient SCA algorithm. Two examples including a sparse component analysis example demonstrate the proposed algorithm.

[1]  Joel A. Tropp,et al.  Greed is good: algorithmic results for sparse approximation , 2004, IEEE Transactions on Information Theory.

[2]  S. Shankar Sastry,et al.  Generalized principal component analysis (GPCA) , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[3]  M. Elad,et al.  $rm K$-SVD: An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation , 2006, IEEE Transactions on Signal Processing.

[4]  Andrzej Cichocki,et al.  Adaptive Blind Signal and Image Processing - Learning Algorithms and Applications , 2002 .

[5]  Michael Zibulevsky,et al.  Underdetermined blind source separation using sparse representations , 2001, Signal Process..

[6]  Allan Kardec Barros,et al.  Independent Component Analysis and Blind Source Separation , 2007, Signal Processing.

[7]  Barak A. Pearlmutter,et al.  Blind Source Separation by Sparse Decomposition in a Signal Dictionary , 2001, Neural Computation.

[8]  Yuanqing Li,et al.  Analysis of Sparse Representation and Blind Source Separation , 2004, Neural Computation.

[9]  Fabian J. Theis,et al.  Sparse component analysis and blind source separation of underdetermined mixtures , 2005, IEEE Transactions on Neural Networks.

[10]  Barak A. Pearlmutter,et al.  Blind source separation by sparse decomposition , 2000, SPIE Defense + Commercial Sensing.

[11]  Andrzej Cichocki,et al.  On-Line K-PLANE Clustering Learning Algorithm for Sparse Comopnent Analysis , 2006, 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings.

[12]  A. Cichocki,et al.  Multilayer nonnegative matrix factorisation , 2006 .

[13]  Mark A. Girolami,et al.  A Variational Method for Learning Sparse and Overcomplete Representations , 2001, Neural Computation.

[14]  Paul S. Bradley,et al.  k-Plane Clustering , 2000, J. Glob. Optim..

[15]  Andrew W. Moore,et al.  Detecting Significant Multidimensional Spatial Clusters , 2004, NIPS.

[16]  Zhaoshui He,et al.  K-EVD Clustering and Its Applications to Sparse Component Analysis , 2006, ICA.

[17]  Terrence J. Sejnowski,et al.  Learning Overcomplete Representations , 2000, Neural Computation.

[18]  Jacob Goldberger,et al.  Hierarchical Clustering of a Mixture Model , 2004, NIPS.

[19]  Rémi Gribonval,et al.  Sparse representations in unions of bases , 2003, IEEE Trans. Inf. Theory.

[20]  Pando G. Georgiev,et al.  Clustering on subspaces and sparse representation of signals , 2005, 48th Midwest Symposium on Circuits and Systems, 2005..

[21]  Kazunori Yamaguchi,et al.  Linear Multilayer Independent Component Analysis for Large Natural Scenes , 2004, NIPS.