Matthew suggested that we can fix the (e.g.) Yamanaya membership to be 100% to one cluster. Together with our modification of using fobs, this will (hopefully) fix the problem we saw in the simulated and real data of the Yamanaya not fully clustering together.
kkdey
commented
8 years ago
Matthew suggested that we can fix the (e.g.) Yamanaya membership to be 100% to one cluster. Together with our modification of using fobs, this will (hopefully) fix the problem we saw in the simulated and real data of the Yamanaya not fully clustering together.