Comparing loo for two complex multivariate models with large data

Interfaces brms
7

Thanks, I’ve made a reproducible example that replicates the error from this post (Moment matching failed), although I haven’t been able to replicate the invalid model error. But this is essentially what I’m trying to do:

# R version 4.0.2
library(loo)#version 2.3.1.9000
library(brms)#version 2.13.9
library(dplyr)

df <- mtcars
df <- mutate(df,ID=c(1:nrow(df)))
f1 = brmsformula('mpg~1+wt+cyl+hp+(1+wt+cyl+hp|ID)')
f2 = brmsformula('qsec~1+wt+cyl+hp+(1+wt+cyl+hp|ID)')
f3 = brmsformula('mpg~1+wt+cyl+(1+wt+cyl|ID)')
f4 = brmsformula('qsec~1+wt+cyl+(1+wt+cyl|ID)')
model_of_interest = brm(f1+f2,data=df,iter=6000,chains=4,cores=4,family="gaussian",control=list(adapt_delta=0.9),save_all_pars=TRUE)
null_model = brm(f3+f4,data=df,iter=6000,chains=4,cores=4,family="gaussian",control=list(adapt_delta=0.9),save_all_pars=TRUE)

loo1 <- brms::loo(model_of_interest, moment_match=TRUE)
loo2 <- brms::loo(null_model, moment_match=TRUE)
1 Like