Has any serious attempt at such cluster analysis ever been made? Where's that mass of data, and where are the papers on this effort, for ME/CFS, or for GWI? If we are now, decades on, still at the point where in both cases the approach is still to look at the whole cohort as a massive, heterogeneous blob, is that situation not a devastating indictment on those who've taken principle responsibility for investigating these illnesses?
Part of any classification or clustering process is to have a defined feature space. Hence a precursor to clustering on a the whole cohort would be to have a consistant way to describe symptoms. This would mean doing things like defining a more accurate language to describe fatigue and other symptoms.
It does seem strange to me that no one has done this especially given it would be easy enough to use standard questionaires across multiple treatment centers, scan them then there is database software that will do much of the analytics.