This repository provides codes to implement the proposed algorithms (group LNGCA and non-Gaussian subspace dimension test/estimation) and reproduce the experiment results in paper "Group Linear non-Gaussian Component Analysis with Applications to Neuroimaging".
The group LNGCA algorithm, located in file fun_call_group_function.R, concists of three steps (1) Subject level LNGCA (2) Group level PCA/SVD (3) Group level ICA/LNGCA. The first step and third step are achieved using codes supporting the single subject LNGCA "Linear non-gaussian component analysis via maximum likelihood", which are included in file fun_LNGCA.R.
We include the implementation, in file fun_dimension_test.R, for (1) the resampling test for a specific hypothesis (2) the dimension estimation algorithm through binary search using dimension test.
The file sim_group_comp_extractoin.R provides all codes to replicate our simulation results.