meeting 2024 02 22 n315 - JacobPilawa/TriaxSchwarzschild_wiki_5 GitHub Wiki
- I've run the second rejection-sampled grid since we've switched to the masked outer kinematics. We discussed the exact points over email and on the previous bullet point.
- Things are still coming along nice, and I think we might be closing in on the true minimum (assuming we don't think that the current Ts/shapes/angles are anything strange).
- Takeaways from this new grid:
- The mass parameters all seem rather well behaved with clear chi2 minima.
- The shapes are a shakier than the mass parameters but I still think they look nice. The minimum in T is something like ~0.9, with Tmaj having a minimum near ~0.03 and Tmin seems to really prefer Tmin = 0.
- First, here are some 1d panels:
| Grid D Alone | Grid CD All Scales | Grid CD Best Scales |
|---|---|---|
 |
 |
 |
For reference, here's the Grid D scaled models plotted in angle space rather than Ts
| Plot |
|---|
 |
- And some statistics for the two grids we have which build this full set of models:
| Plot |
|---|
 |
- I've processed the combined Grid C + D best scaled models with GPR + dynesty. At this stage, we're getting quite consistent results between these cases. I've included a "production" style plot for one of the cases below for reference:
| Cutoff | nu=0.5 | nu=1.5 |
|---|---|---|
| K=40 |  |
 |
| K=60 |  |
 |
| K=80 |  |
 |
| K=100 |  |
 |
| K=150 |  |
 |
- Production Plot:
| Plot |
|---|
 |
- I ran our rejection sampling routine on the "best scaled" version of our models for reference. Here's what the resulting distribution looks like:
| Label | Plot |
|---|---|
| Kinematics |  |
| Beta |  |
Large heatmap of the chi2 per bin/moment
| Plot |
|---|
 |