meeting 2024 02 20 n315 - JacobPilawa/TriaxSchwarzschild_wiki_5 GitHub Wiki
Context
-
We've recently switched to using the set of kinematics with the outer 4 Mitchell data points masked. I recently re-minimized the Grid C models with these updated kinematics, and I've generated proposed set of next points to work with for these kinematics.
-
I think I've landed on the final set/something I'm happy with using our rejection sampling routine -- these were constructed by running GPR on the set of all scaled models, and then sampling points based on a K = 100 cutoff.
-
In the past, we've opted to use the set of best scalings rather than the full scalings for the rejection sampling routine, but I think we have too few models for this to be reliable at this stage. Specifically, when I run the rejection sampling routine on the set of best-scaled models, the resulting Rho0 distribution is far too high/the best scaled models seem to lead to extrapolation to high rho0. I've included the set of points I am proposing below compared to what a typical best-scaled version of points looks like.
Diagnostics
- First, here are the two sets of posteriors which are used for the rejection sampling.
| All Scales [my preference] | Best Scales |
|---|---|
| [images/240220/240220_allpoints_masked_K100_nu1.5-1.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/240220_bestpoints_masked_K100_nu1.5-1.png) |
- And here's what the proposed point distribution would look like for these two cases:
| Parmaeter | All Scales [my preference] | Best Scales |
|---|---|---|
| BH | [images/240220/bh_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/bh_bestpoints_masked_K100_nu1.5_cut100.png) | |
| ML | [images/240220/ml_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/ml_bestpoints_masked_K100_nu1.5_cut100.png) | |
| Rho0 | [images/240220/rho0_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/rho0_bestpoints_masked_K100_nu1.5_cut100.png) | |
| T | [images/240220/T_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/T_bestpoints_masked_K100_nu1.5_cut100.png) | |
| Tmaj | [images/240220/Tmaj_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/Tmaj_bestpoints_masked_K100_nu1.5_cut100.png) | |
| Tmin | [images/240220/Tmin_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/Tmin_bestpoints_masked_K100_nu1.5_cut100.png) | |
| theta | [images/240220/theta_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/theta_bestpoints_masked_K100_nu1.5_cut100.png) | |
| phi | [images/240220/phi_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/phi_bestpoints_masked_K100_nu1.5_cut100.png) | |
| psi | [images/240220/psi_allpoints_masked_K100_nu1.5_cut100.png]]](/JacobPilawa/TriaxSchwarzschild_wiki_5/wiki/[[images/240220/psi_bestpoints_masked_K100_nu1.5_cut100.png) |