Oh my! I just scanned the pdf before heading to bed which may turn out to be a huge mistake if I don't get any sleep tonight. This looks absolutely incredible! Thanks for the head's up.
--- In [email protected], "Mike" <sfclimb...@...> wrote: > > You can look at the sensitivity sections in the IO documentation, found in > the Files section. Individual results are expressed relative to the optimal > combination, and a cumulative distribution is produced. > > Mike > > --- In [email protected], "bh.hicks" <bh.hicks@> wrote: > > > > bump > > > > --- In [email protected], "bh.hicks" <bh.hicks@> wrote: > > > > > > Does anyone have an effective solution for plotting the results of > > > 3-variable optimizations? > > > > > > Probably some type of 3d scatter plot would work the best where you are > > > conceptually filling a volume rather than constructing a surface. I > > > think the ideal solution would literally be a 3d point cloud where every > > > point was color coded and/or sized based on their value and there was a > > > way you could "clip" all the points that fell below a certain threshold > > > so you were only left with the "dense" area somewhere within your volume > > > that represented the optimal variables. > > > > > > It would also be very helpful for potentially identify over-fitting > > > depending on the scale and shape of the "dense" area as you ideally want > > > a nice large mass of good variables rather than just couple of points. > > > > > >
