Previously, the weight used for each line when combining CEL abundances from individual lines was (Uncertainty/Flux)**-2. However, if the user were to input zero as an uncertainty, as they might if they don't have actual uncertainties to put it, this scheme gives infinite weights and Infs and NaNs start appearing.
So I've changed the weight to be just the intensity of the line. I'm opening this issue because there might be better schemes to use than this, and my change is just a bug fix rather than a definitive decision on what the best weighting scheme is.
I've just made a change to this; see
https://github.com/rwesson/NEAT/commit/89e397d50b6a013ef15213b562d531338bcc119c
Previously, the weight used for each line when combining CEL abundances from individual lines was (Uncertainty/Flux)**-2. However, if the user were to input zero as an uncertainty, as they might if they don't have actual uncertainties to put it, this scheme gives infinite weights and Infs and NaNs start appearing.
So I've changed the weight to be just the intensity of the line. I'm opening this issue because there might be better schemes to use than this, and my change is just a bug fix rather than a definitive decision on what the best weighting scheme is.