Operating System: Windows 10 R v. 3.6.0 brms v. 2.15.0 loo v. 2.4.1 I have several GAMMs and would like to choose the best model for my data based on LOO information criterion. However, many of the models have pareto-k values that are too high, as indicated by the loo function. I am concerned th…

Determine number of parameters in brms GAMM to compare to p_loo value

paul.buerkner August 12, 2021, 5:09pm 5

I agree that this could be a reasonable choice indeed.

1 Like

Topic		Replies	Views
Large Pareto K Values in Cognitive Models Modeling rstan , specification , loo , brms	2	147	July 18, 2024
Pareto k values are too high when running model comparison Modeling loo	10	3334	March 11, 2020
A quick note what I infer from p_loo and Pareto k values Modeling loo	35	16495	August 21, 2022
Effective number of parameters for a simple Weibull modell Modeling fitting-issues , loo	3	1837	November 23, 2022
Pareto values versus ELPD differences Modeling loo	5	1074	June 15, 2021