Timeline for Computing or sampling from a posterior with samples observed through a dimensional reduction transformation
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 2 at 10:10 | vote | accept | heckman | ||
| Aug 21 at 7:00 | history | edited | Xi'an | CC BY-SA 4.0 |
added 494 characters in body
|
| Aug 20 at 14:31 | comment | added | Xi'an | @GuillaumeDehaene: This paper is simply an illustration of the principle. If $g$ is of another type, the reconstitution of the sample given the $g(x_i)$'s must be explored anew. If this is not achievable, I would suggest directly running an ABC based on the $g(x_i)$'s, that is simulating new samples such that the generated $g(x_i)$'s are close to the observed $g(x_i)$'s, with or without an MCMC structure. | |
| Aug 20 at 14:28 | history | edited | Xi'an | CC BY-SA 4.0 |
added 112 characters in body
|
| Aug 20 at 14:25 | comment | added | Guillaume Dehaene | And if g is not related to the quantiles / order statistics which you make heavy use if your paper as far as I can tell, would you suggest to use an ABC-like step or do you have another possibility in mind? | |
| Aug 20 at 8:23 | history | answered | Xi'an | CC BY-SA 4.0 |