Second-order Approximation of Minimum Discrimination Information in Independent Component Analysis
This work addresses a specific bottleneck in independent component analysis for signal processing applications, but it is incremental as it builds upon existing FastICA methods.
The paper tackled the performance degradation of FastICA when using multiple nonlinear functions for negentropy estimation by proposing a novel method based on the second-order approximation of minimum discrimination information, which improved efficiency in experiments compared to other ICA algorithms.
Independent Component Analysis (ICA) is intended to recover the mutually independent sources from their linear mixtures, and F astICA is one of the most successful ICA algorithms. Although it seems reasonable to improve the performance of F astICA by introducing more nonlinear functions to the negentropy estimation, the original fixed-point method (approximate Newton method) in F astICA 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.