-
Type:
Task
-
Status: Open (View Workflow)
-
Priority:
Normal
-
Resolution: Unresolved
-
Labels:None
Created on 2023-10-18 09:54:53 by Didier Vibert. % Done: 0
There is a coded zprior that penalize the redshifts for which the solution contains less than 2 lines above SNR 3.5
This prior is not activated (hard coded) and has no dedicated parameter in the parameters.json (like the other z priors activation and strength) .
Shall we enable it and add the corresponding parameter, or remove the lines of code ?
According to me this is not strictly a "prior" in the sense that is based on observations, ie model fitting of the data. The likelihood (least-square term) of the posterior is supposed to handle all the attachment to the data term, where as the prior should only contain a priori information on the parameters...
Nevertheless, even is the term "prior" is improper, we may find that it is useful to enhance z solutions where lines are strong enough. This is more an ad hoc modification of the likelihood.