logLik = 0 throughout analysis

[Sander de Backer]

Dear all,

My first analysis had ESS(N_shifts) > 200 but ESS(logLik) = 0.

I've followed https://bamm-project.org/troubleshooting.html#analysis to troubleshoot and adjust the control file. 

However, every analysis, no matter what I change in the control file, results in ESS(logLik) = 0 throughout the entire analysis.

Does this reflect a problem with my dataset or with the settings in the control file?

I've added the control and run_info files in attachment. Priors were calculated with BAMMtools setBAMMpriors. Here is also some of the terminal output:

  generation    N_shifts    logPrior      logLik   eventRate  acceptRate
           0           0    -13.5037           0          10           1
     1000000           2    -40.0031           0     2.53479     0.71876
     2000000           2    -37.0067           0    0.822629     0.70582
     3000000           1    -27.4315           0     2.61856     0.72569
     4000000           1    -30.7422           0     3.72755    0.718113
     5000000           5    -74.6566           0       5.529    0.712595
  .......................................................................................................

   118000000           1    -39.2435           0    0.602589     0.72149
   119000000           3    -50.6861           0     5.37213    0.714436
   120000000           4     -59.437           0     8.93318    0.718215
   121000000           7     -99.183           0     6.20173    0.715537

Kind regards,

Sander

[Sander de Backer]

I've discovered the problem.

In the control file, sampleFromPriorOnly was set to 0, meaning sampling from prior only and no calculation of likelihoods.

Set sampleFromPriorOnly to 1 and the bamm analysis calculates logLik!

Op woensdag 13 september 2023 om 10:50:54 UTC+2 schreef Sander de Backer: