It remains unclear if the package would also be intended for use to compare a range of different priors at the same time. To smooth such an approach a function would be required which allowed to specify which priors to consider and then have an object where the fitted g, posterors and log-likleihoods might be easily compared to assess which prior might be the most suitable.
We simply need to allow plot to take more than one "ebnm" object. We could plot posteriors on the same graph and also display logliks in the legend. Remember to check that they were produced using the same data; otherwise error out.
Editor's comment:
We simply need to allow plot to take more than one "ebnm" object. We could plot posteriors on the same graph and also display logliks in the legend. Remember to check that they were produced using the same data; otherwise error out.