Hi all, Suppose I have 100 simultaneous time series, what's the best statistical procedure to figure out a transformation of the data, and see if we could squeeze most of the information into a few transformed time series? Thanks!