A new approach for independent component analysis and its application for clustering the economic data

In conventional independent component analysis (ICA) algorithms, the definition of the objective function is typically based on specific dependency criteria. The choice of these criteria significantly influences the performance of the algorithm. This article introduces a general class of dependency criteria, which is based on the cumulative distribution function, to characterise the independence of two variables. Furthermore, an applicable ICA algorithm, grounded in this class and utilising a non-parametric estimator, is proposed. The performance of the proposed algorithm is evaluated and compared with several well-known traditional algorithms, using Amari error estimation calculation as a benchmark. The proposed algorithms have been applied to a real-time series data, serving as a pre-processing clustering method.