search

mjvakili / ccppabc

Approximate Bayesian Computation
MIT License
1 stars 1 forks source link

tex: Referee Report #2 #15

Closed changhoonhahn closed 7 years ago

changhoonhahn commented 7 years ago

if the Gaussian hypothesis underlying the standard approach are so unrealistic as stated in the text, why results are still consistent with the ABC ones? how likely it is that we will face a real data scenario which might lead to significantly different results?

changhoonhahn commented 7 years ago

I think that this comment can be addressed in the discussion by emphasizing that the consistency we find in our comparison is not validation of the Gaussian LF assumption. Instead, it results from the fact that our comparison is apparently not the most sensitive to it. Smaller scales and larger scales (as MJ and Hogg mentioned) -- both of which are very likely in "real data scenarios" -- will likely be more discerning.

changhoonhahn commented 7 years ago

see Issue #17

mjvakili commented 7 years ago

@changhoonhahn : I'm happy with the edits. Closing this issue.