ENA: Spectral index map goodness-of-fit (Phase D deferred) #36
Description
Summary
'Another thing that seems important to have is a "goodness of fit" metric in map form, so that we can visually see how good the non-linear regression fit is on a per pixel basis. Whether it be R-squared or Chi-squared (Bob recommends Chi-Squared), doesn’t matter to me as long as there’s a way to tell how good the fit is.' (Nick Dutton). In this context, we should also add an uncertainty for the spectral index scalar – there is one for the index, just not for the scalar – and document how these uncertainties are derived.
Requirements
- Uncertainty on scalar
- Map variable for goodness-of-fit
Related tickets
None
Estimate
Not estimated