Output¶
The structure of the PTArcade output is the following:
- By default,
out_dir = ./chains
. The user can specify a different output directory via the configuration file. - By default,
name=np_model
. The user can specify a different name in the model file.
The output generated by PTArcade matches that of ENTERPRISE and it includes, beyond the MC chains, a number of files which summarize key information on the run and the MC sampler. For our purposes, the most important files produced by PTArcade are:
pars.txt
-
This file contains the names of the model parameters. The order in which the parameters appear in this file will also dictate the order in which the parameters appear in the
chain_1.txt
file.When running with
mode="ceffyl"
, thepars.txt
file for the example model discussed in the introduction will read as follows: (1)- When running with
mode = "enterprise"
, in addition to the user-specified parameters,pars.txt
will also include intrinsic red noise parameters (two per pulsar) and, in the case of the IPTA dataset, DM parameters.
- When running with
chain_1.txt
-
This file contains the MC chains. It is formatted such that each line represents an MC sample and each column corresponds to a parameter of our model. The ordering of the parameters, i.e., which column is associated to each parameter, can be read out from the
pars.txt
file. (1)- Note that, when running with
mode = "enterprise"
, in addition to the user-specified parameters, the chains will also include intrinsic red noise parameters (two per pulsar) and, in the case of the IPTA dataset, DM parameters.
In addition to the model parameters, the last four columns of each row report the values of the posterior, the likelihood, the acceptance rate, and an indicator variable for parallel tempering, which does not matter in our case since PTArcade does not use parallel tempering.
For the example model discussed in the introduction, the output of a run in Ceffyl-mode will be:
Here, the first two columns give the values of \(\log_{10}A_*\) and \(\log_{10}(f_*/{\textrm Hz})\) while the remaining columns give the value of the posterior, the likelihood, the acceptance rate, and the parallel-tempering indicator.
- Note that, when running with
priors.txt
-
The prior file is similar to the
pars.txt
file, but, in addition to the parameter names, it also includes their prior distributions. Here is an example for our test model, when running in Ceffyl-mode: