Nonlinear Weighted Independent Component Analysis

Andrzej Bedychaj, Przemyslaw Spurek, Aleksandra Nowak, Jacek Tabor

Abstract: Independent Component Analysis (ICA) aims to find a coordinate system in which the components of the data are independent. In this paper we construct a new nonlinear ICA model, called WICA, which obtains better and more stable results than other algorithms. A crucial tool is given by a new efficient method of verifying nonlinear dependence with the use of computation of correlation coefficients for normally weighted data. In addition, we propose a new nonlinear mixing to perform baseline comparison experiments and a reliable measure that allows for a fair evaluation of nonlinear models. Code for methods presented in the paper is available on our GitHub.

External IDs:dblp:conf/ipmu/BedychajS0T22