Timeline for What is an "uninformative prior"? Can we ever have one with truly no information?
Current License: CC BY-SA 4.0
21 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 23, 2020 at 11:32 | history | edited | Xi'an | CC BY-SA 4.0 |
edited body
|
| Aug 7, 2020 at 3:41 | history | edited | Xi'an | CC BY-SA 4.0 |
added 8 characters in body
|
| May 29, 2020 at 20:05 | history | edited | Xi'an | CC BY-SA 4.0 |
added 40 characters in body
|
| Jan 3, 2018 at 21:30 | history | edited | Xi'an | CC BY-SA 3.0 |
added 582 characters in body
|
| Nov 11, 2015 at 4:44 | history | bounty ended | Glen_b | ||
| Nov 9, 2015 at 13:10 | history | edited | Xi'an | CC BY-SA 3.0 |
added 235 characters in body
|
| Nov 7, 2015 at 15:23 | comment | added | Xi'an | Correct: The Fisher information is derived from the density behind the observations, thus uses a frequency property of the distribution. Most parameters are also defined in terms of (frequency) moments of this distribution. Whether this is "good" or "bad" remains a matter of opinion. | |
| Nov 7, 2015 at 15:15 | history | edited | Xi'an | CC BY-SA 3.0 |
added 90 characters in body
|
| Nov 7, 2015 at 11:29 | comment | added | gwr | Yes, I thought about that, but that amounts a bit to frequentism in itself from my preliminary understanding, doesn't it? Somehow MaxEnt seemed clearer from what I know which definitely is a kind of information that needs maximization... ;) | |
| Nov 7, 2015 at 10:31 | comment | added | Xi'an | @gwr: if you read the reference prior literature, you will see that they consider the expected information, so it does not depend on the data. | |
| Nov 7, 2015 at 9:58 | comment | added | gwr | Regarding maximizing the information brought in by the data: Why should the prior be dependent upon the data or the likelihood for that? | |
| Nov 7, 2015 at 9:33 | history | edited | Xi'an | CC BY-SA 3.0 |
added 613 characters in body
|
| Nov 7, 2015 at 9:28 | history | edited | Xi'an | CC BY-SA 3.0 |
added 613 characters in body
|
| Jun 7, 2015 at 12:44 | comment | added | P.Windridge | Re. uniform prior- Olle Haggstrom gives a very nice talk on this ("Notes From My Personal Crusade Against The Uniform Distribution"). The uniform distribution can say something utterly ludicrous, it is a modelling assumption, and definitely not the same as no information! | |
| Jan 13, 2012 at 19:35 | comment | added | Manoel Galdino | It's almost unfair. After all, he's Christian Robert! Just kidding. Great answer. And I'd love if @Xi'an could expand it in a post at his blog, specially about how parametrization is important to the topic of "uninformative" priors. | |
| Jan 3, 2012 at 18:26 | vote | accept | Fomite | ||
| Jan 3, 2012 at 18:26 | comment | added | Fomite | An outstanding answer. Thank you. And yet another book to go on the wish list. | |
| Jan 3, 2012 at 15:00 | comment | added | whuber♦ | +1 Thank you for a good and well-informed overview of the issues. | |
| Jan 3, 2012 at 15:00 | comment | added | cardinal | (+1) For an objective (no less!), straightforward answer. | |
| Jan 3, 2012 at 13:07 | comment | added | Elvis | (+1) Your book? Oh damn. I so have 387 questions for you :) | |
| Jan 3, 2012 at 11:50 | history | answered | Xi'an | CC BY-SA 3.0 |