Calculate Training Error for Machine Learning Models

Computes the training error for a given machine learning model using resampled datasets. This function is designed to fit the model to each split of the data, predict outcomes, and calculate a performance metric (e.g., R-squared) for each resample.

Usage

get_training_error(model, splits, metric)

Arguments

model

A machine learning model object compatible with the fit_resamples method.

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 performance metric (R-squared by default) for each resample and an identification of the error type as "Training".

See also

fit_resamples, mutate

Examples

if (FALSE) { # \dontrun{
library(tidymodels) # assuming tidymodels includes necessary packages
data(iris)
model <- linear_reg() |> set_engine("lm")
splits <- initial_split(iris, prop = 0.75)
training_error <- get_training_error(model, splits)
} # }