initializations: Computes number of initializations to be performed by ica_pca
Description
For a given number of sub-Gaussian and super-Gaussian sources
and Gaussian components, the function ica_pca will initialize
the model multiple times. This function will compute the number
of initializations that will be performed.
If the number of initializations is small (less than 50 to 100),
the ica_pca function may fail to identify the optimal model;
models with small numbers of initializations should be run
several times using different values for seed and/or
offset_random. As the number of sources and components gets
large (e.g., with totals more than 10) the number of
initializations grows quickly. To a first approximation,
computation time is proportional to the number of
initializations.