-
Notifications
You must be signed in to change notification settings - Fork 9
Altering SuperPlot appearance
To affect the changes described below, either restart SuperPlot, or simply click Make plot.; the Make plot. button reloads the SuperPlot modules.
To alter the labels for the axes, either
-
Alter the ParamNames.py file - if the names list is the same length as the data in your chain, these names are used by default.
-
Alter your *info file - this file is used only if the names list in ParamNames.py is the wrong length.
The colour schemes and styles of the SuperPlot are in Appearance.py; each data item (e.g. posterior mean, posterior pdf) has an associated object which holds its colour scheme, labels etc.
The Appearance.py also contains the default plot title and legend titles, and other plot options, which you might wish to change.
The posterior pdf and profile likelihood are found by histogramming the data in the chain, and integrating the posterior weights and maximizing the likelihood over bins in this histogram, respectively. The result is sensitive to the number of bins per dimension, and the bin limits - i.e. whether data outside the interval of the bin limits is discarded.
By default, no bin limits are applied - the histogram spans the full range of the data, and 70 bins per dimension are used. These settings can be altered in Appearance.py - the default setting might be undesirable if outliers ruin the binning.