scikit-learn-contrib/fastcan

A fast canonical-correlation-based search algorithm for feature selection, system identification, data pruning, etc.