distribution: MAP from KDE or just from sample histogram
lukasheinrich opened this issue · comments
If MAP from histogram, depends strongly on binning. becomes an issue, if binning too fine for statistics available for posterior (directly related to how much time we want to spend in emcee)
