Incorporating the Temporal Structure of Sources into Independent Component Analysis
Seungjin Choi · Journal of Electrical Engineering and Information Science · 1999
Independent component analysis (ICA) is a statistical method, the goal of which is to decompose sensor signals (observations) into a linear sum of statistically independent components. Most existing methods for ICA have not taken the temporal structure of sources into account. This paper presents an ICA method which incorporates with the temporal structure of sources, hence the resulting algorithm is efficient for the separation of mixtures of temporally correlated sources. Useful behavior of the proposed method is demonstrated through computer simulations.