-
Notifications
You must be signed in to change notification settings - Fork 71
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Do not add a noisemodel to a Model by default anymore #667
Comments
A similar warning as to Pandas could be given:
|
One question we need to answer: what do we do with the parameter uncertainty, for example the standard error in the fit report, when noise is set to False? The uncertainty will likely be underestimated, as there will be much autocorrelation in the residuals. There are several options:
I will start a branch to make some of the needed changes, so we can use this for testing. |
Based on proposal from @raoulcollenteur in issue #667
Good question. I think I am in favor of not displaying the standard error anymore whether the noise model is used or not. After all, just using it doesn't necessarily mean the autocorrelation is gone and so the standard error may still not be representative. Then we should probably create a separate function to compute the standard error. |
Comment from @mbakker7 in PR #678 (comment):
|
As discussed at the meeting of 2024-4-9, we will not add a noisemodel to a Model object by default anymore. If a noisemodel is required, the user can add it using In Pastas 2.0 we will also remove the |
A discussed in the last Pastas meeting, we plan to make
noise=False
the new default for ml.solve for Pastas 2.0 release. This Issue is to document the work related to this big change. Here are some of the reasons why we plan to make this change:To do list related to this Issue:
[ ] Investigate alternative methods for uncertainty quantification with autocorrelated errors.
[ ] Document this change throughout pastas
[ ] Provide FutureWarnings in the last version before 2.0 (whenever that is).
[ ] ...?
The text was updated successfully, but these errors were encountered: