generalize to arbitrary dimensions
bhomass opened this issue · 2 comments
bhomass commented
This code works for 2 dimensions. Can it be expanded to work for arbitrary number of dimensions, like over 100?
bhomass commented
Apparently, it would be excessively expensive in computation load when this method is extended further.
syrte commented
Agreed. The computation cost will be too high for ND KS test, as it uses quadrants. Also it is unclear how well it works for higher dimensions.
The energy distance statistics test may work. Maybe you can check with literature.