search

choderalab / LiquidBenchmark

GNU General Public License v2.0
0 stars 6 forks source link

Labeling of outliers in Figs 1 and 2? #18

Open jchodera opened 9 years ago

jchodera commented 9 years ago

Would be great if we could automatically have outliers in Figs 1 and 2 labeled, or find some other way to represent them.

kyleabeauchamp commented 9 years ago

To my eye, the figures don't really seem to have a problem with outliers. For example, the uncorrected version of the dielectrics have some points that are "far off", but they are fixed by adding back in the missing physics.

Or do you just want to add stars next to points that are sufficiently far from y=x to at some pvalue threshold?

jchodera commented 9 years ago

Today, I was specifically asked which compounds were the ones with the largest relative deviations in the density plot, so I thought others might be interested in that information as well. The supplementary info has all of the plots, but given that the axes are different for each plot and the titles are tiny, it's actually really hard to figure that out.