predict_gam {tidymv} | R Documentation |
Get predictions from a GAM model.
Description
This function is from the superseded package tidymv. Please, use the tidygam package instead.
It returns a tibble with the predictions from all the terms in a gam or bam model.
If you simply want to return a tibble with the predicted values of the
response/outcome variable based on all terms (minus excluded smooth terms),
set type = "link"
(the default). Note that if type = "link"
,
parametric terms cannot be excluded from the prediction, due to limitations
of mgcv
. If you want to return a tibble with the predicted values of
the response/outcome variable for each term in the model separately, set
type = "terms"
. This type can be helpful if you want more flexibility
in plotting.
Usage
predict_gam(
model,
exclude_terms = NULL,
length_out = 50,
values = NULL,
type = "link"
)
Arguments
model |
A |
exclude_terms |
Terms to be excluded from the prediction. Term names should be given as they appear in the model summary (for example, |
length_out |
An integer indicating how many values along the numeric predictors to use for predicting the outcome term (the default is |
values |
User supplied values for specific terms as a named list. If the value is |
type |
Either |
Value
A tibble with predictions from a gam or bam model.
Examples
## Not run:
library(mgcv)
set.seed(10)
data <- gamSim(4)
model <- gam(y ~ fac + s(x2) + s(x2, by = fac) + s(x0), data = data)
# get predictions
p <- predict_gam(model)
# get predictions excluding x0 (the coefficient of x0 is set to 0);
# setting the value for the excluded term to NULL with the argument 'values'
# reduces computation time
p_2 <- predict_gam(model, exclude_terms = "s(x0)", values = list(x0 = NULL))
# get predictions with chosen values of x0
p_3 <- predict_gam(model, values = list(x0 = c(0.250599, 0.503313, 0.756028)))
## End(Not run)