nuclear-multimessenger-astronomy / nmma

A pythonic library for probing nuclear physics and cosmology with multimessenger analysis
https://nuclear-multimessenger-astronomy.github.io/nmma/
GNU General Public License v3.0
30 stars 58 forks source link

light_curve_analysis for multiple models #214

Closed tylerbarna closed 7 months ago

tylerbarna commented 1 year ago

Please fill out relevant sections below; remove those that are unused.

Feature Summary A function on top of the existing light_curve_analysis that allows for multiple models to be analyzed at once

Usage / behavior Call multi_model_analysis with the same arguments as light_curve_analysis, but including multiple models and priors in the arguments will result in analyses being executed for each model. Once all analyses are concluded, an optional plot flag will plot all of the model best fits together with the data.

Alternative Solutions Could alter the behaviour of light_curve_analysis to have adding multiple models and priors to do what I described rather than doing a mix of the models, but I think the suggestion of another function would be better for compatibility

Implementation details Additional command line function that just calls light_curve_analysis for each associated model and prior before a unique plotting feature to plot them all together onto one plot (with corner plots remaining separate)

Additional context Would need to account for models taking different amounts of time to fit and also implement some type of cutoff to account for models that stall. Will also need to account for the potential of failed fits.

sahiljhawar commented 1 year ago

I think multi-model analysis already exists. I don’t exactly remember but somewhere @tsunhopang mentioned that the fluxes/luminosities add up if multiple models are being used at once.

mcoughlin commented 1 year ago

@sahiljhawar I think what @tylerbarna wants is a function that will loop over model configurations, and presumably compute odds ratios for you.

sahiljhawar commented 1 year ago

Hmm okay, maybe I understood it wrong.

tylerbarna commented 1 year ago

@sahiljhawar I think what @tylerbarna wants is a function that will loop over model configurations, and presumably compute odds ratios for you.

yeah, this is more along the lines of what I was thinking; the existing ability to use multiple models as a combined fit is why I suggested a separate command to avoid confusion/compatibility issues