Exhaustively expands the combinations of `features`. The number of formulas is `choose(n, k)` summed over the sizes, which grows quickly (20 features, sizes 2 to 20, is over a million), so the call stops above `max_formulas`.
Usage
expand_feature_formulas(
features,
k = NULL,
min_terms = 2,
max_terms = length(features),
response = NULL,
max_formulas = 1e+05
)Arguments
- features
Character vector of terms.
- k
One size, or `NULL` for the range `min_terms` to `max_terms`.
- min_terms, max_terms
Size range when `k` is `NULL`.
- response
Optional left-hand side as text, for example `"Surv(time, status)"` or `"outcome"`; a trailing `~` is accepted.
- max_formulas
Safety limit on the number of formulas.
Value
A tibble with `id`, `formula` (list of formulas), `terms` (list of character vectors) and `n_terms`, ordered by size and then by the order of `features`.
Examples
expand_feature_formulas(c("a", "b", "c"), k = 2, response = "y ~")
#> # A tibble: 3 × 4
#> id formula terms n_terms
#> <int> <list> <list> <int>
#> 1 1 <formula> <chr [2]> 2
#> 2 2 <formula> <chr [2]> 2
#> 3 3 <formula> <chr [2]> 2
expand_feature_formulas(c("a", "b", "c"))
#> # A tibble: 4 × 4
#> id formula terms n_terms
#> <int> <list> <list> <int>
#> 1 1 <formula> <chr [2]> 2
#> 2 2 <formula> <chr [2]> 2
#> 3 3 <formula> <chr [2]> 2
#> 4 4 <formula> <chr [3]> 3