pws3141/clusterICA: Approximate Independent Component Anaylsis

An implementation of approximate Independent Component Analysis, using clustering and optimisation of random projections of the data. It combines ordering of random points on the projective space and clustering of the best points using divisive k-means until some tolerance condition is met, with an optimisation step for each cluster. Each loading is found sequentially, and Householder rotations are used to make each loading orthogonal to all previous ones.

Getting started

Package details

MaintainerPaul Smith <>
Package repositoryView on GitHub
Installation Install the latest version of this package by entering the following in R:
pws3141/clusterICA documentation built on July 14, 2020, 5:04 a.m.