MULTIDIMENSIONAL INDEPENDENT COMPONENT ANALYSIS FOR STATISTICAL ESTIMATIONS OF INDETERMINACIES IN ELECTROCARDIOGRAMS

Independent component analysis (ICA) is a technique capable of separating independent components (ICs) from complex electrocardiogram (ECG) signals. The basic intention behind using multidimensional independent component analysis (MICA) is to find stable higher dimensional source signal subspaces. This study highlights the ability of ICA for parametrization of ECG signals to reduce the amount of redundant ECG data if any in a data set. The aim of this paper is to justify the underlying theory of the use of ICA and how it can be extended to for MICA separation of the ECG signals for combinational leads to attain most useful diagnostic information, which was not discussed in other some similar previous publications in this field. It is also investigated that the value of kurtosis coefficients for the ICs, which represents the noise component, can be further reduced using parametrized multidimensional independent component analysis (PMICA) technique. The indeterminacies available in the ECG data are also analyzed using modified version of Jade algorithm for PMICA and parametrized standard independent component analysis (PSICA). For the ECG data set, Jade algorithm is applied first to find smaller subspaces for MICA analysis and can therefore be regarded as a basis algorithm for PMICA analysis. The simulation results are obtained in Matlab environment to indicate that, ICA can definitely improve signal-noise ratio (SNR) in minimizing the reconstruction errors. The future scope of MICA expected by author is that, by reconsidering the notion of ICA, a more general perspective can be envisioned: i.e. modified multidimensional independent component analysis (MMICA). It would be based on a morphological geometric parametrization (MGP) which would further reduce the indeterminacies involved in matrix-based modeling (MBM).