ICA Algorithm: Implementation with Maximum Likelihood and BFGS Optimization

This document presents the Independent Component Analysis (ICA) algorithm, which is equivalent to Bell and Sejnowski's 1995 Infomax approach [1] formulated using maximum likelihood estimation. The implementation assumes no noise and requires the number of observations to equal the number of sources. Optimization is performed using the BFGS method [2], with dimensionality reduction via PCA and independent component count determination using Bayes Information Criterion (BIC) [3].

Steepest Descent Method in Nonlinear Programming with MATLAB Implementation

MATLAB source codes for nonlinear programming methods including Steepest Descent, BFGS method, and Conjugate Gradient method (featuring 3-Quasi-Newton BFGS method implementation)