Computes the validation error for a given machine learning model using resampled datasets.
This function fits the model to each split of the data and then collects and filters the performance metrics, specifically focusing on R-squared (rsq) for validation.
Usage
get_validation_error(model, splits, metric)
Arguments
- model
A machine learning model object, expected to be compatible with the fit_resamples method from the tune package.
- splits
An object containing data splits, typically generated by functions from the
rsample package, used for resampling.
- metric
The performance metric of interest as a string.
Value
A data frame containing the R-squared metric for each resample along with an identification of the error type as "Validation".
Examples
if (FALSE) { # \dontrun{
library(dplyr)
library(tidymodels) # assuming tidymodels includes necessary packages
data(iris)
model <- linear_reg() %>% set_engine("lm")
splits <- vfold_cv(iris, v = 5)
validation_error <- get_validation_error(model, splits)
} # }
