Evaluate a tuning spec

Usage

tuning_eval(
  tuning,
  n,
  kx,
  k,
  se = FALSE,
  jackknife = "none",
  ctx = NULL,
  mode = c("default", "grid")
)

Arguments

tuning: Tuning spec or registered name.
n: Integer sample size.
kx: Integer localization covariate dimension.
k: Integer DNN neighbor count.
se: Logical; whether jackknife SEs are requested.
jackknife: One of "fast", "loo", or "none".
ctx: Optional context passed to tuning evaluators.
mode: One of "default" or "grid".

Value

mode = "default": list with s, meta, and feasibility.
mode = "grid": list with candidates data.frame and meta.

Examples

tuning_eval("paper_undersmooth", n = 500, kx = 2, k = 1)
#> $s
#> [1] 42
#> 
#> $meta
#> $meta$lambda
#> [1] 0.4
#> 
#> $meta$formula
#> [1] "lambda * n^((kx + 1)/(kx + 2))"
#> 
#> $meta$rounding
#> [1] "floor"
#> 
#> $meta$name
#> [1] "paper_undersmooth"
#> 
#> $meta$n
#> [1] 500
#> 
#> $meta$kx
#> [1] 2
#> 
#> 
#> $feasibility
#> $feasibility$k_le_s
#> [1] TRUE
#> 
#> $feasibility$s_lt_n
#> [1] TRUE
#> 
#> 
tuning_eval("paper_undersmooth", n = 500, kx = 2, k = 1, mode = "grid")
#> $candidates
#>   lambda  s feasible reason
#> 1   0.30 31     TRUE     ok
#> 2   0.40 42     TRUE     ok
#> 3   0.50 52     TRUE     ok
#> 4   0.75 79     TRUE     ok
#> 
#> $meta
#> $meta$name
#> [1] "paper_undersmooth"
#> 
#> $meta$n
#> [1] 500
#> 
#> $meta$kx
#> [1] 2
#> 
#>