Kernel selection with evolutionary algorithm for multiple kernel independent component analysis

Kernel independent component analysis (KICA) has an important application in blind source separation, in which how to select the optimal kernel, including the kernel functional form and its parameters, is the key issue for obtaining the optimal performance. In practices, a single kernel is usually chosen as the kernel model of KICA in light of experience. However, selecting a suitable kernel model is a more difficult problem if one has not sufficient experience. To deal with this problem, an evolution based method to select the kernel model of KICA is proposed in this paper. There are two main features of the proposed method: one is that using a multiple kernel model, a convex combination of several single kernels, replaces the single kernel model; another is that particle swarm optimization (PSO) algorithm is utilized to find the combination weights of the composite kernel. Experiments conducted on separating one-dimensional mixed signals, nature images, and spectroscopic CCD images showed that using multiple kernels model with PSO kernel selection algorithm can enhance the performance of KICA.

[1]  Marco Wiering,et al.  2011 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN) , 2011, IJCNN 2011.

[2]  Riccardo Poli,et al.  Particle swarm optimization , 1995, Swarm Intelligence.

[3]  K. Horne,et al.  AN OPTIMAL EXTRACTION ALGORITHM FOR CCD SPECTROSCOPY. , 1986 .

[4]  Wang Xing-yu,et al.  KICA Model Selection Method with Application to Removing Artifact from EEG , 2006 .

[5]  Licai Deng,et al.  LAMOST Spectral Survey , 2012, 1206.3569.

[6]  Scott C. Douglas,et al.  Self-stabilized gradient algorithms for blind source separation with orthogonality constraints , 2000, IEEE Trans. Neural Networks Learn. Syst..

[7]  Nello Cristianini,et al.  Learning the Kernel Matrix with Semidefinite Programming , 2002, J. Mach. Learn. Res..

[8]  Walid Hachem,et al.  A CLT FOR INFORMATION-THEORETIC STATISTICS OF NON-CENTERED GRAM RANDOM MATRICES , 2011, 1107.0145.

[9]  Aapo Hyvärinen,et al.  Independent component analysis: recent advances , 2013, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[10]  Michael I. Jordan,et al.  Kernel independent component analysis , 2003, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03)..

[11]  Jian Yu,et al.  Adaptive regularization deconvolution extraction algorithm for spectral signal processing , 2014, 2014 IEEE International Symposium on Innovations in Intelligent Systems and Applications (INISTA) Proceedings.

[12]  Ethem Alpaydin,et al.  Multiple Kernel Learning Algorithms , 2011, J. Mach. Learn. Res..

[13]  D. Alter.  PUBLICATIONS OF THE ASTRONOMICAL SOCIETY OF THE PACIFIC , 2016 .

[14]  Peng Wu,et al.  A pre-selecting base kernel method in multiple kernel learning , 2015, Neurocomputing.

[15]  Peter L. Bartlett,et al.  Classification with a Reject Option using a Hinge Loss , 2008, J. Mach. Learn. Res..

[16]  Peng Wu,et al.  Method of Evolving Non-stationary Multiple Kernel Learning , 2014, ICONIP.

[17]  Anhua Chen,et al.  Kernel Function and Parameters Optimization in KICA for Rolling Bearing Fault Diagnosis , 2013, J. Networks.

[18]  Gunnar Rätsch,et al.  Large Scale Multiple Kernel Learning , 2006, J. Mach. Learn. Res..

[19]  Zhongfu Ye,et al.  A Sky-Subtraction Algorithm for LAMOST Using Two-Dimensional Sky-Background Modeling , 2012, Publications of the Astronomical Society of Australia.