Currently QU-fitting reports the median of the posterior distributions as the 'best fit' value. This is fine for single component cases, but is problematic for multi-component fits which are often multi-modal.
I'm currently investigating a solution using a KDE to find the modal value, but I thought I should raise the issue for book-keeping purposes.
Currently QU-fitting reports the median of the posterior distributions as the 'best fit' value. This is fine for single component cases, but is problematic for multi-component fits which are often multi-modal.
I'm currently investigating a solution using a KDE to find the modal value, but I thought I should raise the issue for book-keeping purposes.