Second-Order Approximation of Minimum Discrimination Information in Independent Component Analysis
YunPeng Li · IEEE Signal Processing Letters · 2021
Independent Component Analysis (ICA) is intended to recover the mutually independent sources from their linear mixtures, and$FastICA$is one of the most successful ICA algorithms. Although it seems reasonable to improve the performance of$FastICA$by introducing more nonlinear functions to the negentropy estimation, the original fixed-point method (approximate Newton method) in$FastICA$degenerates under this circumstance. To alleviate this problem, we propose a novel method based on the second-order approximation of minimum discrimination information (MDI). The joint maximization in our method is consisted of minimizing single weighted least squares and seeking unmixing matrix by the fixed-point method. Experimental results validate its efficiency compared with other popular ICA algorithms.