Transpose output lpd of proj_linpred()?

[fweber144]

This reprex:

library(projpred)
library(rstanarm)
options(mc.cores = parallel::detectCores(logical = FALSE))
data("df_gaussian", package = "projpred")
mydat <- cbind("y" = df_gaussian$y, as.data.frame(df_gaussian$x))
myfit <- stan_glm(y ~ V1 + V2 + V3 + V4 + V5,
                  family = gaussian(),
                  data = mydat,
                  prior = hs(df = 1, global_scale = 0.01),
                  seed = 1140350788)

mylinpred <- proj_linpred(myfit,
                          solution_terms = c("V3", "V2"),
                          newdata = mydat[1:3, , drop = FALSE],
                          ndraws = 20)
str(mylinpred)

shows that (in the default case of integrated = FALSE) either output element mylinpred$pred or output element mylinpred$lpd should probably be transposed (I guess it should be element mylinpred$lpd to match the behavior of proj_predict() and also to match rstan's convention of having as many rows as posterior draws).

[AlejandroCatalina]

I have transposed both pred and lpd in proj_linpred.

[fweber144]

Only one of them needed to be transposed; as mentioned above:

either output element mylinpred$pred or output element mylinpred$lpd should probably be transposed (I guess it should be element mylinpred$lpd to match the behavior of proj_predict() and also to match rstan's convention of having as many rows as posterior draws).

[AlejandroCatalina]

Right, I addressed this on latest develop (still to be uploaded). Best, Alejandro

…

On 6. Apr 2021, 18.12 +0300, Frank Weber ***@***.***>, wrote: Only one of them needed to be transposed; as mentioned above: > either output element mylinpred$pred or output element mylinpred$lpd should probably be transposed (I guess it should be element mylinpred$lpd to match the behavior of proj_predict() and also to match rstan's convention of having as many rows as posterior draws). — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.