Closed jwang-lilly closed 2 years ago
Hi Jian,
the order is not alphabetical. The elements of opt_cut$boot
belong to opt_cut$subgroup
. With the example data opt_cut$boot[[1]]
is for the female group:
library(cutpointr)
library(tidyverse)
cutpointr(data = suicide,
x = dsi,
class = suicide,
subgroup = gender,
boot_runs = 100,
method = maximize_metric,
metric = sum_sens_spec,
use_midpoints = TRUE) %>%
select(subgroup, optimal_cutpoint, roc_curve, boot)
#> Assuming the positive class is yes
#> Assuming the positive class has higher x values
#> Running bootstrap...
#> # A tibble: 2 x 4
#> subgroup optimal_cutpoint roc_curve boot
#> <chr> <dbl> <list> <list>
#> 1 female 1.5 <roc_cutpointr [11 x 10]> <tibble [100 x 23]>
#> 2 male 2.5 <roc_cutpointr [11 x 10]> <tibble [100 x 23]>
Created on 2022-02-26 by the reprex package (v2.0.1)
P.S. Thanks for the kind words! Glad that you like the package.
Excellent Christian. Thanks so much!
Hi Christian,
Thanks much for the nice and fast package to make cupoint related calculations enjoyable!
I have a quick question. When using subgroup option in the cutpointr(), for example subgroup = 'gender', the opt_cut$boot would be a list of two items. How can I tell which item for gender = female and which is for gender = male? I guess it is based on alphabetical order therefore female in opt_cut$boot[[1]] and male in opt_cut$boot[[2]], is this correct?
Thanks a lot.
@Thie1e