Hi everybody,
I'm using brms package in R for logistic regression. I compared two models using the 'looic' command. The output contains the following information:
Output of model 'SA1nest.brm1':
Computed from 4000 by 91 log-likelihood matrix
Estimate SE
elpd_loo -54.9 8.4
p_loo 15.3 3.7
looic 109.9 16.7
------
Monte Carlo SE of elpd_loo is NA.
Pareto k diagnostic values:
Count Pct. Min. n_eff
(-Inf, 0.5] (good) 86 94.5% 920
(0.5, 0.7] (ok) 4 4.4% 221
(0.7, 1] (bad) 0 0.0%
(1, Inf) (very bad) 1 1.1% 12
See help('pareto-k-diagnostic') for details.
Output of model 'SA1nest.brm2':
Computed from 4000 by 91 log-likelihood matrix
Estimate SE
elpd_loo -48.3 5.8
p_loo 6.5 1.3
looic 96.6 11.7
------
Monte Carlo SE of elpd_loo is 0.1.
All Pareto k estimates are good (k < 0.5).
See help('pareto-k-diagnostic') for details.
Model comparisons:
elpd_diff se_diff
SA1nest.brm2 0.0 0.0
SA1nest.brm1 -6.7 4.2
............................................
Now I understand that the model with the least looic value is the best. I just wanted to know what the negative value implies under the Model comparisons. Somebody please elaborate.
Thanks