Performance Analysis Based on Transform Matrix of MLE in CICA

This paper proposes a new method for the cost function based on the maximum likelihood estimation (MLE) in the complex valued ICA problem. The method from complex field to real field is very valid to construct cost function in the complex valued ICA. Simulation experiments with two sub-Gaussian signals as source signals are carried out under the framework, and the performance evaluation of four kinds of the transform matrixes are analysis. Based on the results of the experiments, the orthogonal matrix as the transform matrix from complex field to real field to resolve the complex valued ICA problem is best in the four kinds of matrixes of the paper.