predict.model_fit  R Documentation 
Apply a model to create different types of predictions.
predict()
can be used for all types of models and uses the
"type" argument for more specificity.
## S3 method for class 'model_fit' predict(object, new_data, type = NULL, opts = list(), ...) ## S3 method for class 'model_fit' predict_raw(object, new_data, opts = list(), ...) predict_raw(object, ...)
object 
An object of class 
new_data 
A rectangular data object, such as a data frame. 
type 
A single character value or 
opts 
A list of optional arguments to the underlying
predict function that will be used when 
... 
Arguments to the underlying model's prediction
function cannot be passed here (see

If "type" is not supplied to predict()
, then a choice
is made:
type = "numeric"
for regression models,
type = "class"
for classification, and
type = "time"
for censored regression.
predict()
is designed to provide a tidy result (see "Value"
section below) in a tibble output format.
When using type = "conf_int"
and type = "pred_int"
, the options
level
and std_error
can be used. The latter is a logical for an
extra column of standard error values (if available).
For censored regression, a numeric vector for time
is required when
survival or hazard probabilities are requested. Also, when
type = "linear_pred"
, censored regression models will by default be
formatted such that the linear predictor increases with time. This may
have the opposite sign as what the underlying model's predict()
method
produces. Set increasing = FALSE
to suppress this behavior.
With the exception of type = "raw"
, the results of
predict.model_fit()
will be a tibble as many rows in the output
as there are rows in new_data
and the column names will be
predictable.
For numeric results with a single outcome, the tibble will have
a .pred
column and .pred_Yname
for multivariate results.
For hard class predictions, the column is named .pred_class
and, when type = "prob"
, the columns are .pred_classlevel
.
type = "conf_int"
and type = "pred_int"
return tibbles with
columns .pred_lower
and .pred_upper
with an attribute for
the confidence level. In the case where intervals can be
produces for class probabilities (or other nonscalar outputs),
the columns will be named .pred_lower_classlevel
and so on.
Quantile predictions return a tibble with a column .pred
, which is
a listcolumn. Each list element contains a tibble with columns
.pred
and .quantile
(and perhaps other columns).
Using type = "raw"
with predict.model_fit()
will return
the unadulterated results of the prediction function.
For censored regression:
type = "time"
produces a column .pred_time
.
type = "hazard"
results in a list column .pred
containing tibbles
with a column .pred_hazard
.
type = "survival"
results in a list column .pred
containing tibbles
with a .pred_survival
column.
In the case of Sparkbased models, since table columns cannot contain dots, the same convention is used except 1) no dots appear in names and 2) vectors are never returned but typespecific prediction functions.
When the model fit failed and the error was captured, the
predict()
function will return the same structure as above but
filled with missing values. This does not currently work for
multivariate models.
library(dplyr) lm_model < linear_reg() %>% set_engine("lm") %>% fit(mpg ~ ., data = mtcars %>% dplyr::slice(11:32)) pred_cars < mtcars %>% dplyr::slice(1:10) %>% dplyr::select(mpg) predict(lm_model, pred_cars) predict( lm_model, pred_cars, type = "conf_int", level = 0.90 ) predict( lm_model, pred_cars, type = "raw", opts = list(type = "terms") )
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.