I have this set of data which can be regularly spaced if needed. The problem is that there are different regimes and the boundaries are non-rectangular.
For example, there is a change of phase that happens at the y = 1 - 1/x, meaning that the data behaves differently above and below this graph. I can sample the data everywhere, but I would like the extrapolation to be different above and below this line, as there would be problems extrapolating along the boundary.
At the moment, I can only think of cutting the spaces in rectangles, but that seems a bit complicated to do.
If anyone has any ideas of how I could do it best, that would be great ^^.
Thanks!
