Interpretation of the results of Leave-One-Out (LOO) cross-validation

When I run the code of LOO, I got the following results. But I do not know the means of these results. So what are 'elpd_loo', 'p_loo', 'looic', 'elpd_diff' represent for? In this result, which model is better? Thank you!

fit1 <- stan_glm(mpg ~ wt, data = mtcars, refresh = 0)
fit2 <- stan_glm(mpg ~ ht + cyl, data = mtcars, refresh = 0)
loo1 <- lPreformatted textoo(fit1, cores = 2)
print(loo1)

     Estimate    SE

elpd_loo -15339.7 140.6
p_loo 25.6 2.5
looic 30679.3 281.1


Monte Carlo SE of elpd_loo is 0.1.
All Pareto k estimates are good (k < 0.5).

loo2 <- loo(fit2, cores = 2)

print(loo2)

     Estimate    SE

elpd_loo -15339.1 140.1
p_loo 23.5 2.2
looic 30678.2 280.3


Monte Carlo SE of elpd_loo is 0.1.
All Pareto k estimates are good (k < 0.5).

loo_compare(list(loo3, loo4))
elpd_diff se_diff
fit2 0.0 0.0
fit1 -0.5 1.9

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.