Hi Candy,
So the general idea is to interpolate in a transformed space, i.e.
Yes, we could add that. I guess log might be a common use case though, so we could also consider adding a specialized
Thanks,
Christoph
So the general idea is to interpolate in a transformed space, i.e.
f^-1(interp(x, f(y))
?Yes, we could add that. I guess log might be a common use case though, so we could also consider adding a specialized
LogTransformedInterpolation
wrapper that is hard-coded to Log/Exp but could also support derivatives out of the box (if the inner interpolation supports it), even though we would loose integration support. What do you think?Thanks,
Christoph