Handle known variability for certain types of stars

[karllark]

There are types of stars that vary intrinsically. AGB stars, Cepheids, etc. Often different bands are taken at different times folding causing the an "uncertainty" in the fit not captured in the current physics models. A model uncertainty could be added to account for this.

[s-goldman]

I've talked to a bunch of statisticians and SED-fitting people about fitting multi-epoch data for variable stars and it doesn't look like there is a graceful way of doing it. So long as there is enough data, the "median" SED may be useful, but I'm not sure how that can be done with our current framework.

[marthaboyer]

We know typical amplitudes at different wavelengths and we know typical periods, so can't we model the uncertainties? And correlate bands that are simultaneous? (Sorry I'm missing hack day!)

[karllark]

@s-goldman : Can you provide more details of the reasons why handling the variability is not easy/possible? The case I'm thinking about is when we only have 1 measurement - so no median is possible.

[s-goldman]

Fitting each epoch is fine, it only seems to be a problem when you fit the combined epochs.