If you:
then formulize
may be for you. Formulize is very new, but you can install it from CRAN if you have R 3.4+ with:
install.packages("formulize")
You can get the development version (recommended) with:
# install.packages("devtools")
devtools::install_github("alexpghayes/formulize")
Suppose you want to add a formula interface to an existing modelling function, say cv.glmnet
. Then you could do the following
library(recipes)
library(glmnet)
library(formulize)
glmnet_cv <- formulize(cv.glmnet)
glmnet_model <- glmnet_cv(mpg ~ drat + hp - 1, mtcars)
predict(glmnet_model, head(mtcars))
#> 1
#> Mazda RX4 22.44822
#> Mazda RX4 Wag 22.44822
#> Datsun 710 22.95815
#> Hornet 4 Drive 19.94218
#> Hornet Sportabout 17.62212
#> Valiant 19.15914
Similarly glmnet_cv
works with recipe objects like so
rec <- recipe(mpg ~ drat + hp, data = mtcars)
glmnet_model2 <- glmnet_cv(rec, mtcars)
predict(glmnet_model2, head(mtcars))
#> 1
#> [1,] 22.35392
#> [2,] 22.35392
#> [3,] 22.85062
#> [4,] 19.97897
#> [5,] 17.72884
#> [6,] 19.24084
You may also be interested in the more dangerous exciting version genericize
, which you should call for its side effects.
genericize(cv.glmnet)
form <- mpg ~ drat + hp - 1
X <- model.matrix(form, mtcars)
y <- mtcars$mpg
set.seed(27)
mat_model <- cv.glmnet(X, y, intercept = TRUE)
set.seed(27)
frm_model <- cv.glmnet(form, mtcars, intercept = TRUE)
set.seed(27)
rec_model <- cv.glmnet(rec, mtcars, intercept = TRUE)
predict(mat_model, head(X))
#> 1
#> Mazda RX4 22.25028
#> Mazda RX4 Wag 22.25028
#> Datsun 710 22.73249
#> Hornet 4 Drive 20.01959
#> Hornet Sportabout 17.84620
#> Valiant 19.33092
predict(frm_model, head(mtcars))
#> 1
#> Mazda RX4 22.25028
#> Mazda RX4 Wag 22.25028
#> Datsun 710 22.73249
#> Hornet 4 Drive 20.01959
#> Hornet Sportabout 17.84620
#> Valiant 19.33092
predict(rec_model, head(mtcars))
#> 1
#> [1,] 22.25035
#> [2,] 22.25035
#> [3,] 22.73255
#> [4,] 20.01946
#> [5,] 17.84608
#> [6,] 19.33070
This creates a new S3 generic cv.glmnet
, sets the provided function as the default method (cv.glmnet.default
), and adds methods cv.glmnet.formula
and cv.glmnet.recipe
using formulize
.
This will mask cv.glmnet
and features no safety checks because safety isn’t fun.
formulize
doesn’t do anything special with intercepts. This means that you need to careful with functions that require you to specify intercepts in non-standard ways, such as cv.glmnet
above.formulize
will probably break.glmnet
, take a look at glmnetUtils.