CosmoMC minimizer likelihoods

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
Elizabeth Gould
Posts: 13
Joined: January 30 2015
Affiliation: Queen's University

CosmoMC minimizer likelihoods

Post by Elizabeth Gould » February 10 2016

In the .minimum file produced by the minimizer, the sum of the \chi^2 values is less than the total \chi^2 at the top of the file. I would like to know where the difference is coming from (and what the \chi^2 should be).

For instance, one of the files from the PLANCK team:
(base_plikHM_TT_lowTEB_lensing_post_BAO.minimum
from http://www.cosmos.esa.int/web/planck/pla)
-log(Like) = 5638.37102200327
chi-sq = 11276.7420440065

4.620 9.240 CMB: lensing = smica_g30_ftl_full_pp
5247.428 10494.857 CMB: lowTEB = lowl_SMW_70_dx11d_2014_10_03_v5c_Ap
383.100 766.200 CMB: plik = plik_dx11dr2_HM_v18_TT
0.005 0.010 BAO: 6DF
0.704 1.407 BAO: MGS
1.201 2.402 BAO: DR11CMASS
0.240 0.479 BAO: DR11LOWZ
The sum of \chi^2 values is 11274.595.

My run with similar datasets:
-log(Like) = 5659.86416122493
chi-sq = 11319.7283224499

21.419 42.839 CMB: BKPLANCK = BKPlanck_detset_comb_dust
4.650 9.300 CMB: lensing = smica_g30_ftl_full_pp
383.066 766.131 CMB: plik = plik_dx11dr2_HM_v18_TT
5247.442 10494.885 CMB: lowTEB = lowl_SMW_70_dx11d_2014_10_03_v5c_Ap
0.003 0.006 BAO: 6DF
0.736 1.473 BAO: MGS
1.201 2.402 BAO: DR11CMASS
0.212 0.423 BAO: DR11LOWZ
The sum of \chi^2 values here is 11317.459.

Antony Lewis
Posts: 1652
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: CosmoMC minimizer likelihoods

Post by Antony Lewis » February 10 2016

It is probably the contribution from priors, c.f.

https://github.com/cmbant/CosmoMC/pull/1

Elizabeth Gould
Posts: 13
Joined: January 30 2015
Affiliation: Queen's University

CosmoMC minimizer likelihoods

Post by Elizabeth Gould » February 10 2016

Thanks.

Post Reply