search

IvanNikolic21 / Lyman-alpha-bubbles

Codes used to derive the likelihood calculation for finding an ionized bubble given some Lyman-emitter data
0 stars 0 forks source link

Out-of-distribution tests -> gaussian distribution for mocks #57

Open IvanNikolic21 opened 2 months ago

IvanNikolic21 commented 2 months ago

What if I use a different distribution for the equivalent widths compared to the fiducial one. This is the test I'm running right now and I'll put the results here.

IvanNikolic21 commented 2 months ago

The test has passed, there is basically no difference:

Image

This is the result for the gaussian distribution

IvanNikolic21 commented 2 months ago

Image

Still marginally worse inference, but fully consistent with what I was finding before.

IvanNikolic21 commented 2 months ago

Final result (for now, more runs will follow):

Image

The inference is worse, but not significantly. Note here that I'm changing the quantiles of the distribution which I'll start to do for other runs.