tlienart
commented
7 years ago - Should also specify that it works with truncated distributions
- In CCL should argue that while Bayesian seems inappropriate for n>>p, it may be appropriate for cases where n ~ p and lots of structure such as a graphical model ...