Timeline for Why not report the mean of a bootstrap distribution?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 3, 2013 at 21:21 | history | edited | David J. Harris | CC BY-SA 3.0 |
parenthesis
|
| Sep 29, 2013 at 19:59 | comment | added | Momo | Thanks, that's exactly what I thought as well. To me bagging does not seem to make much sense for anything other than the estimate of a response, so it is limited in that sense. | |
| Sep 29, 2013 at 19:41 | comment | added | David J. Harris | I normally see bagging used to reduce the variance of one's estimates for the response (i.e. its sensitivity to fluctuations in the data). The most commonly bagged models (e.g. trees) typically don't have well-defined parameters that would be easily compared across bootstrap samples. | |
| Sep 29, 2013 at 18:04 | comment | added | David J. Harris | I'm not sure I understand your question. I didn't actually use the phrase "bias of a parameter." I'm also not entirely sure what you mean by $y$ versus $\theta$ in this context. | |
| Sep 29, 2013 at 3:43 | history | answered | David J. Harris | CC BY-SA 3.0 |