search

sofia-calgaro / ZeroNuFit.jl

Bayesian unbinned fit of the neutrinoless double-beta decay
MIT License
1 stars 1 forks source link

Parameter names #7

Closed tdixon97 closed 1 week ago

tdixon97 commented 2 weeks ago

Currently some systematic uncertainity parameters (resolution and bias) have indexs. Perhaps its better to change them to have names?

sofia-calgaro commented 1 week ago

this seems to work in likelihood.jl, but when accessing the params later on (eg for plotting) it still gives "bias" and "res" as names, not 𝛥 or σ:

(S = 3.471476909924879e-27, B = 0.00033980281804643175, α = 1.2299859462479807, res = [1.269495619868158, 1.252753755769896, 1.1066306669943076, 1.067997715868386, 1.0897944462005074, 1.4276863581032984, 1.4736882006806298, 1.0876186919118982, 1.4918312967879053, 1.1466640337151894, 1.2032327691677212, 1.6911038013542319, 1.1388967824659328], bias = [-0.15189034802590287, -0.39187568457752636, -0.05832044065867748, 0.047443906241417144, -0.02898869383079352, -0.031559273836899805, -0.1013333742088224, -0.1642461046335973, -0.1974499125252953, 0.03175023645375571, 0.01992325875931887, -0.5496944363827646, -0.2431152798550923])

tdixon97 commented 1 week ago

I wouldnt close this yet since we still refer to resolution and bias as \sigma[i] and \Delta[i] and i thought a name would be clearer and avoid bugs like instead \sigma[:part_001_B0000B] instead

tdixon97 commented 1 week ago

The Dict of par names for plots mostly resolves this issue.. maybe we can just give partitions name (for plots and its enough?)