search

Chris-Pedersen / LaCE_manager

Cosmological emulator for the 1D flux power spectrum of the Lyman-alpha forest
3 stars 1 forks source link

Implement joint analysis of different P1D measurements #51

Closed andreufont closed 2 years ago

andreufont commented 2 years ago

This could be trivial if we were careful with the coding, but it might be very convoluted if we were not. Hard to tell.

Ideally, we would just write a "combined likelihood" object that multiplies the likelihood of both measurements (say DESI mock + HIRES mock). Or we would add an option to the emcee sampler to have two likelihoods, and add their loglike before adding the prior.

I'd be happy to look at it myself, but I thought @Chris-Pedersen might have useful inputs.

andreufont commented 2 years ago

Actually, the easiest strategy might be to define a combined_data, and leave the likelihood unchanged.

andreufont commented 2 years ago

This was implemented in #54