Discussion of “Random-projection ensemble classification" by Timothy I. Cannings and Richard J. Samworth