A Version of the FastICA Algorithm Based on the Secant Method Combined with Simple Iterations Method
The work proposes a new algorithm for the estimation of the ICA model, an algorithm based on secant method and successive approximations. The first sections briefly present the standard FastICA algorithm based on the Newton method and a new version of the FastICA algorithm. The proposed algorithm to estimate the independent components combines the secant iterations with successive approximations technique. The final section presents the results of a comparative analysis experimentally derived conclusions concerning the performance of the proposed method. The tests were performed of several samples of signal files.
KeywordsIndependent Component Analysis Blind Source Separation Numerical Method
- 2.Cho, Y.H., Hong, S.J., Park, S.M.: Face Recognition by Using Combined Method of Centroid Shift and Fixed-Point ICA Based on Secant Method. In: Proceedings of the Sixth International Conference on Intelligent Systems Design and Applications (ISDA 2006), IEEE Computer Society, Los Alamitos (2006)Google Scholar
- 3.State, L., Cocianu, C., Panayiotis, V., Constantin, D.: PRINCIPAL DIRECTIONS - BASED ALGORTHM FOR CLASSIFICATION TASKS. In: Proceedings of the Ninth International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2007), Timisoara, September 26-29, 2007, IEEE Computer Society, Los Alamitos (2007)Google Scholar
- 4.Stone, J.V.: Independent Component Analysis. The MIT Press, Cambridge (2004)Google Scholar
- 5.Bartlett, M.S., Movellan, J.R., Sejnowski, T.J.: Face Recognition by Independent Component Analysis. IEEE Transactions on Neural Networks 13 (2002)Google Scholar