Timeline for Loss function for conditional variance?
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 25, 2023 at 16:25 | answer | added | picky_porpoise | timeline score: 1 | |
| Sep 27, 2022 at 18:53 | comment | added | Richard Hardy | Somewhat related: "What problem or game are variance and standard deviation optimal solutions for?". | |
| Aug 19, 2022 at 4:55 | comment | added | Cagdas Ozgenc | ARMA-GARCH or a similar mean-variance combo (estimating mean and variance simultaneously) estimator works much better in practice and more efficient. Almost all of those models are estimated by maximum likelihood (which is also an M-estimator by the way). Maybe you are trying to ask something else? | |
| Aug 19, 2022 at 4:48 | comment | added | Cagdas Ozgenc | In my answer to my own question I had to assume a normal distribution and a mean (or a conditional mean) of zero. This is in contrast to say squared loss which works for any distribution with a defined mean. More interestingly squared loss is not the only function that yields the mean. There are more efficient ones depending on the underlying distribution. Similarly Linex function is actually a terrible estimator for variance. stats.stackexchange.com/questions/442365/…. | |
| Aug 10, 2022 at 14:35 | answer | added | Xiubo Zhang | timeline score: 2 | |
| Aug 10, 2022 at 14:21 | history | asked | Dave | CC BY-SA 4.0 |