Timeline for State of the art algorithms for estimating noise variance
Current License: CC BY-SA 4.0
14 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 8, 2019 at 8:41 | comment | added | Laurent Duval | Since wavelets are zero-average, the highest scale should be zero-mean. Yet indeed, you can use other robust estimators, aggregate patches, etc. I mostly meant to provide the main principles | |
| Jan 7, 2019 at 23:17 | comment | added | user14717 | @LaurentDuval: Does the derivation of the MAD estimator assume that the mean of the highest scale subband is close to zero? Otherwise you would expect to use the median absolute deviation from the median. | |
| Dec 7, 2018 at 10:05 | comment | added | Laurent Duval | Nice feedback. Hence my (too-narrow) focus on methods that tend to separate transients from stationary parts, or that kind of stationarize random processes | |
| Dec 6, 2018 at 22:22 | comment | added | mark leeds | Just to finish: Estimating time-series econometric models, in finance particularly, is quite a tricky and often un-successful endeavor. So, just giving you the heads up that they're not the be-all end-all. | |
| Dec 6, 2018 at 22:21 | comment | added | mark leeds | Laurent: Just a disclaimer because I've done those sorts of things in that paper. They all assume stability-stationarity in the time-series world, whether it's after differencing or without differencing. So, their applicability to the dsp world is going to have the same issues as time-seriess: namely, the question of stationarity. As soon as there's some kind of regime-change, strucural break-, change in underlying mean, they all kind of fall apart because the likelihood is basically wrong. | |
| Dec 5, 2018 at 21:26 | comment | added | Laurent Duval | Thank you @Mark Leeds. This is now pinned on my "to read" wall (which, unlike the Chinese wall, can be seen from Pluto. | |
| Dec 5, 2018 at 7:42 | comment | added | mark leeds | Thanks Laurent for additions. I looked for the Schwert paper but the algorithm seems somewhat ad-hoc so I'm not sure if it's useful even if I found it. The link here will atleast give econometric references that may be of interest. All the best. .drphilipshaw.com/Protected/… | |
| Dec 5, 2018 at 7:24 | comment | added | Laurent Duval | Comments added, and more references as well for images | |
| Dec 5, 2018 at 7:24 | history | edited | Laurent Duval | CC BY-SA 4.0 |
added details and references
|
| Dec 5, 2018 at 0:16 | comment | added | user14717 | Hopefully quick question: Where does the 0.6745 come from? | |
| Dec 5, 2018 at 0:15 | vote | accept | user14717 | ||
| Dec 4, 2018 at 22:53 | comment | added | Laurent Duval | I should check the literature you mention. What seems to matter is the generally of the model (AR, sparse) and the associated degree of freedom | |
| Dec 4, 2018 at 22:45 | comment | added | mark leeds | Hi Laurent: Probability a stupid question but are you saying to investigate using the lags of the response itself ? if so, then that's autoregressive modelling in econometrics and there are ways to decide when to stop addings lags. Shwert ( maybe 1970's) has some paper discussing the method but I forget the title. | |
| Dec 4, 2018 at 22:35 | history | answered | Laurent Duval | CC BY-SA 4.0 |