fit_model() takes a model_spec object from {parsnip} and
it fits is. Then, usual methods like summary(), or coef() can be applied
directly on it, while it can still be used as the {tidymodels} recommends it.
Usage
fit_model(data, formula, ..., type = NULL, env = parent.frame())
# S3 method for model_fit
summary(object, ...)
# S3 method for model_fit
anova(object, ...)
# S3 method for model_fit
plot(x, y, ...)
# S3 method for model_fit
chart(data, ..., type = "model", env = parent.frame())
# S3 method for model_fit
as.function(x, ...)
# S3 method for model_fit
coef(object, ...)
# S3 method for model_fit
vcov(object, ...)
# S3 method for model_fit
confint(object, parm, level = 0.95, ...)
# S3 method for model_fit
fitted(object, ...)
# S3 method for model_fit
residuals(object, ...)
# S3 method for model_fit
rstandard(model, ...)
# S3 method for model_fit
cooks.distance(model, ...)
# S3 method for model_fit
hatvalues(model, ...)
# S3 method for model_fit
deviance(object, ...)
# S3 method for model_fit
AIC(object, ..., k = 2)
# S3 method for model_fit
BIC(object, ...)
# S3 method for model_fit
family(object, ...)
# S3 method for model_fit
nobs(object, ...)
# S3 method for model_fit
formula(x, ...)
# S3 method for model_fit
variable.names(object, ...)
# S3 method for model_fit
labels(object, ...)Arguments
- data
A data frame (or a model_fit object for
chart())- formula
A formula specifying a model
- ...
Further arguments passed to the method
- type
The type of model fitting, specified by a model_spec object or the name of such an object in a string
- env
The environment where to evaluate
type. It isparent.frame()by default and you probably have no reasons to change it, unless you really know what you are doing!- object
A model_fit object
- x
Idem
- y
Not used here
- parm
Specification of parameters for the confidence intervals (vector of numbers or of names). If missing, all parameters are considered.
- level
Confidence level required.
- model
Idem
- k
The penalty per parameter to be used in the AIC (by default,
k = 2).
Examples
library(parsnip)
data(trees, package = "datasets")
# Take the habit to prefix your regression model specs by `reg_`
reg_lm <- linear_reg(mod = "regression", engine = "lm")
trees_fit <- fit_model$reg_lm(data = trees, Volume ~ Girth)
#> Error in x(type = name, ...): The type= argument must provide a model_spec object or its name in a character string.
# You can use summary(), AIC(), anova(), tidy(), glance(), etc. directly
summary(trees_fit)
#> Error in eval(expr, envir, enclos): object 'trees_fit' not found
anova(trees_fit)
#> Error in eval(expr, envir, enclos): object 'trees_fit' not found
AIC(trees_fit)
#> Error in eval(expr, envir, enclos): object 'trees_fit' not found
coef(trees_fit)
#> Error in eval(expr, envir, enclos): object 'trees_fit' not found
library(chart)
chart(trees_fit)
#> Error in eval(expr, envir, enclos): object 'trees_fit' not found
# etc.